0:03 hello and welcome my name is Shannon
0:04 Kemp and I'm the chief digital manager
0:06 for day diversity we want to thank you
0:07 for joining the latest in the monthly
0:09 webinar series lessons in data modeling
0:11 with Donna Burbank and sponsored today
0:14 by idira today Donna will be discussing
0:16 conceptual data models how to get the
0:17 attention of business users just a
0:19 couple of points to get us started due
0:20 to the large number of people that
0:22 attend these sessions you will be muted
0:24 during the webinar we very much
0:25 encourage you to chat with us and with
0:27 each other throughout the webinar to do
0:29 so just click the chat icon on the top
0:31 right hand corner for that feature for
0:32 questions we'll be collecting them via
0:34 the Q&A section in the bottom right hand
0:36 corner of your screen or if You' like to
0:38 tweet we encourage you to share
0:39 highlights or questions via Twitter
0:42 using hashtag lessons DDM as always we
0:44 will send a follow-up email within two
0:45 business days containing links to those
0:47 slides and links to the recording of the
0:49 session and any additional information
0:51 requested throughout the webinar now let
0:52 me introduce to you our speaker for
0:55 today Donna Burbank she is a recognized
0:56 industry expert in Information
0:58 Management with over 20 years of
1:00 experience helping organizations enrich
1:01 their business opportunities through
1:04 data and information she is currently
1:06 manage the managing director of global
1:08 data strategy limited where she assist
1:09 organizations around the globe in
1:12 driving value from their data she has
1:13 worked with dozens of Fortune 500
1:15 companies worldwide in the America's
1:17 Europe Asia and Africa and speaks
1:19 regularly in Industry conferences in
1:21 fact she will be speaking at one of the
1:23 day diversity upcoming conferences
1:26 Enterprise data World 2017 in Atlanta
1:29 the first week of April here and just um
1:31 little over a week she'll be starting
1:33 with a tutorial on best practical steps
1:35 to implementing metadata strategy at
1:38 that event so Donna hello and
1:41 welcome hello thank you always a
1:43 pleasure to do these webinars and it's
1:45 good to see some familiar I would say
1:46 familiar faces but familiar names and
1:49 some some of the names are faces are
1:50 familiar as well so thanks guys whove
1:53 joined us on a regular basis um and as
1:55 Shannon mentioned today we are talking
1:57 about conceptual data models and
1:58 hopefully we'll make them less
2:00 conceptual and F focus more on the
2:02 second part of the conversation which is
2:03 really how to they get the attention of
2:06 business users and probably more even
2:08 valuable um how to get business value
2:11 out of these models so uh Shannon has
2:13 always did an excellent overview of me
2:16 so I don't want to keep going on that um
2:18 just a few things I am on Twitter at
2:21 Donna Burbank and then today there's a
2:24 hashtag lessons DM so if you are a
2:26 Twitter fan uh continue the conversation
2:28 online and I know there's always a lot
2:31 of good questions uh at these events so
2:32 hopefully we'll try to leave some time
2:36 for that at the end um I think U my bio
2:38 is up there so kind of had a long
2:41 background in data modeling metadata
2:43 management both in the field and with
2:45 several uh product vendors written a few
2:47 books on data modeling and probably most
2:50 notable for this conversation um was
2:52 product ma director of product
2:55 management for a while at idira or back
2:58 then it was um in barcadero ER studio so
3:00 any of the really great uh features of
3:02 that product that you use are mine and
3:03 anything that doesn't work you don't
3:06 like was built after I left so just
3:09 kidding so that's a little bit about me
3:12 um a bit about the series and as Shannon
3:14 mentioned these are all recorded um and
3:15 the slides are available too so if you
3:17 missed any in the past they are still
3:19 out there and I know a lot of folks who
3:21 are busy during the day you know often
3:22 have to catch them later so we
3:24 appreciate that as well uh so we've done
3:26 some things on Enterprise architecture
3:28 business intelligence in the past you'll
3:30 see the theme is to really put data
3:32 modeling in context which hopefully
3:33 we'll do a little bit slightly
3:35 differently this time but hopefully in
3:38 that same theme um his data modeling end
3:40 of itself is is well we might find it
3:41 interesting but it's not really that
3:43 interesting as just a standalone
3:45 artifact and that's what we'll try to
3:47 head on today if it's really only useful
3:48 and useful to a business when you're
3:51 doing something else with it like bi or
3:52 understanding the Enterprise and that
3:54 kind of thing uh so that's kind of the
3:56 theme we tried to cover as well as some
3:58 of the new technical Technologies things
4:00 like graph databases that are
4:02 H now so hopefully you can join us in
4:04 some of the future events as well they
4:05 are all available for registration if
4:07 you want to get them on your calendar
4:08 early I know some people do it all in
4:10 January and then just have them out
4:11 there so appreciate those folks that
4:14 seem to keep coming back so I appreciate
4:17 that so today the topic is going to be
4:18 really about how a data model
4:21 facilitates communication and that to me
4:24 um is really the main topic of
4:27 conceptual data modeling um is is that
4:29 business focus so if we remember and
4:31 we'll go more about this the focus is
4:33 really on a business audience and so the
4:34 display and the rules and the things
4:36 we're capturing really should be all
4:38 about the business at this level and the
4:41 second note I love distress um
4:43 Simplicity does not mean lack of
4:45 importance and and I think a lot of us
4:47 in the industry the more simple you make
4:49 it you should have SE get less credit
4:52 right anyone can make it hard um can you
4:55 take a very complex topic like an
4:56 organization or all the data that's
4:58 flowing through an organization and
5:00 really sum that up in a one pager um and
5:02 that's really the beauty of a
5:05 well-formed uh conceptual data model and
5:06 and that that's the goal is to make
5:07 something very complicated and make it
5:09 really intuitive so you can really sus
5:11 out those Core Concepts that really make
5:14 sense um and as always a model or a tool
5:17 or any technology is only part of the
5:19 solution and I think this is even more
5:21 important when we are talking about uh a
5:23 conceptual data model because the whole
5:25 point of it's communication and for some
5:26 of you I might go a little outside the
5:29 lines of what some people think of when
5:30 they think of
5:32 conceptual data model and I'll stick to
5:34 it because I think the goal is
5:36 communication right so if someone wants
5:38 to fly a plane with you know smoke
5:39 signals in the back to get the attention
5:42 of the business um go for it you know I
5:44 think so I think we have to be a little
5:46 more creative sometimes to get some of
5:47 the stakeholders that we're not
5:49 particularly used to dealing with so and
5:51 the process and best practices around
5:53 that to get that consensus and buy in
5:54 you know so that plays into some of the
5:55 other things we'll be talking about
5:56 throughout the year like the had
5:57 governance and some of those things that
6:00 really are another piece of the puddle
6:02 puzzle when we talk about data
6:04 modeling um so again you've probably
6:06 seen this framework again hopefully it's
6:08 helpful to kind of Bring It Back uh to
6:11 some some of the Core Concepts so you
6:12 know in our practice we always start
6:14 with the top down why are we doing this
6:16 what's the business strategy and how
6:18 does that align to a data strategy and
6:19 then we start from the bottom up as well
6:22 of what all the technology and we're
6:23 talking about unstructured data here is
6:26 it all relational databases etc etc and
6:28 all the pieces in between and I don't
6:29 won't go in detail on this at time if
6:31 you've joined other webinars you've seen
6:33 this but what I wanted to cover here is
6:35 really how data modeling and data
6:37 architecture and I'm going to include
6:39 metadata in this because a lot of the
6:41 value in a data model is the metadata
6:43 around it are really a core part of any
6:45 strategy you can have a great strategy
6:46 but if you don't know what your data
6:48 means or or where it is or how it's
6:51 stored um and particularly uh in this
6:52 conversation what it means and and the
6:54 rules around it you're not going to get
6:56 much value so we often start at the very
6:58 top of a business strategy starting with
7:00 a conceptual data model what are the
7:02 Core Business Concepts we're worried
7:04 about we have a campaign we're talking
7:06 about customers prospects product you
7:08 know those are the core pieces of a
7:09 business and I think when you
7:11 communicate that right to a business
7:13 audience folks get that in you're part
7:15 of the conversation and how you really
7:17 can link that
7:19 strategy so again you've probably seen
7:20 this triangle too if you've joined other
7:22 webcast but again hopefully it's helpful
7:24 to put it in context so when I'm talking
7:26 about business data I tend to stay
7:29 either at this conceptual model um or
7:31 you know it does tie into the logical
7:33 and we could have a whole topic just on
7:36 this but you know is a logical model
7:38 physically focused or business focus and
7:40 I'll say both and you can see that line
7:42 is strategically placed it captures
7:44 business rules and at this point you are
7:46 starting to think about how that may
7:48 apply to a data structure the focus is
7:50 still business um but there is some
7:52 structure around that that we're trying
7:53 to put in but I would say with the
7:55 conceptual level we really are really
7:57 trying to model the business unless the
7:59 technology or it shouldn't be at all
8:00 about technology behind it it should be
8:03 completely a business focus um you also
8:05 see that the pyramid gets smaller at the
8:08 top and that sort of hits on the point I
8:10 mentioned before is that less is more
8:12 and simplicity is the beauty of a
8:13 conceptual model so if you can think of
8:15 the bottom of that pyramid where there
8:18 may be thousands of physical tables can
8:20 you logically and conceptually summarize
8:22 those that we're really just getting
8:25 that view of customers product Styles um
8:27 inventory that kind of thing that's the
8:29 beauty of a conceptual model so simp is
8:32 better so you've heard me say conceptual
8:34 modeling we always get the question so
8:36 hopefully proactively we'll say this and
8:37 I will say there's our friend
8:39 Shakespeare up in the corner if you
8:41 didn't recognize him um or I'll say the
8:43 cob children have no sues right as an
8:45 industry the whole point is creating
8:46 common definitions we're pretty bad at
8:49 that ourselves um and there's a lot of
8:50 different names for this model I'm
8:53 talking about uh some people call us
8:55 subject area model some call business
8:58 data model Enterprise model um seems to
9:00 be the major theme uh is is folks call
9:02 that a conceptual data model so in the
9:05 book that I referenced in my bio um data
9:06 modeling for the business my co-author
9:09 Steve hberman actually did this not a
9:11 very scientific survey we had it out in
9:13 at that day the DM review with a big
9:16 publication um of what people call this
9:17 thing we describe this thing without
9:20 using a word we've called it in our book
9:22 a highlevel model just to avoid that
9:24 conversation because in full disclosure
9:25 Steve hobman calls it a subject area
9:28 model so I think those 12% go to his
9:29 classes I'm just kidding a lot of
9:31 respect for Steve he's a good guy so I
9:34 can tease him um and the rest of us call
9:35 it a conceptual data model so I can say
9:38 haha Steve I'm right because we actually
9:40 did a survey but I particularly don't
9:41 care uh what you call it I think the
9:43 point and hopefully by the end of this
9:44 presentation is it doesn't matter what
9:46 you call it you probably shouldn't call
9:47 it a conceptual data model to a business
9:49 user I I would actually recommend not
9:51 doing that that sounds very vague right
9:53 so maybe it's your business model your
9:54 business data model your business
9:56 information about whatever you call it
9:58 um the point is communication so I won't
9:59 argue sotic although I think I just
10:03 spent a minute doing that see I'm guilty
10:05 myself um but we we tend to do that
10:06 because we're modelers and we think of
10:08 semantics and definitions and sometimes
10:10 we can go a little too far so I will
10:12 talk about that more in the presentation
10:14 but if you call the thing I'm discussing
10:16 any of these things that's fine we'll
10:18 have you in our club um and as long as
10:19 you know what I'm talking about I think
10:22 we're fine so highle model I think is
10:23 what we use in the book just to keep it
10:26 neutral um so and just to set the
10:28 context I mentioned things about
10:30 metadata so technical metadata is hugely
10:33 important um often folks start there
10:35 because it's easy and in many cases to
10:36 kind of reverse engineer I won't say
10:38 managing technical Med is necessarily
10:40 easy but um you know it's going to get
10:43 your structure of energ an energer Car
10:45 Bar car whatever we're talking today of
10:47 more of the business metadata so what do
10:49 we mean by employee what do we mean by
10:51 customer and then of course the data is
10:53 the actual customers themselves but um
10:55 more of the reasons I put a picture of
10:57 Mr John there is um we'll come back to
11:00 this I think visual of presentation
11:01 especially at the business level is so
11:04 important um and I'd like to highlight
11:05 that when I build my conceptual models
11:07 because really if you're thinking of a
11:09 business these are concrete things one
11:10 of the reasons I don't like the name
11:12 conceptual data model it's actually to
11:14 me the opposite we're talking about the
11:16 actual concrete things that run a
11:18 business right locations and customers
11:20 and products and employees those are the
11:22 things they're things they they exist
11:24 they live and breed in some cases so um
11:26 kind of remind that in the picture I have
11:27 have
11:31 there uh Med again is this is a survey
11:35 um report uh I did earlier last late
11:38 last year uh with dataversity um on what
11:39 are some of the emerging Trends in
11:41 metadata management again this was
11:43 broadly metadata and not specifically
11:44 data modeling but what was interesting
11:47 is that 80% of the users of metadata are
11:49 the business and and when we did a
11:51 survey of a lot of the customers um or
11:53 users of metadata a lot of the
11:55 conversation was around just this it
11:57 helps business users and it really
11:59 understand the context around
12:01 information um and really the usage and
12:03 the business rules and all of that so
12:05 for those folks who say oh you know that
12:07 the business doesn't care about data I
12:09 would I would argue with you um that I
12:10 think 80% are saying they are I think
12:12 there's some frustration of how we
12:14 communicate to those folks in a way they
12:16 understand and one tool by all means not
12:18 the only Tool uh but one tool at least
12:20 at a starting point can be this
12:21 conceptual data model so that they will
12:24 be interested in listening I do believe
12:26 in fact my next quote I I just remember
12:27 this from one of my Consulting
12:29 engagements we we
12:31 the group try sort of convince the you
12:32 know we were in the IT department
12:34 talking to the business about how we
12:36 have to get common definitions and and
12:38 almost apologizing for talking to them
12:39 of we're trying to get the relationships
12:41 between data and and they said you mean
12:42 you're not doing this
12:44 already they sort of expected that we
12:46 were doing a lot more of this business
12:49 level design in our systems um and and
12:50 we're sort of shocked that we hadn't so
12:51 we were a little embarrassed but that
12:54 was a a positive feedback that of course
12:55 you guys would do this we would want you
12:57 to have these definitions and you know I
12:59 think sometimes so the how how do we
13:03 have this um linkage between folks that
13:04 you know their day job is either selling
13:06 product or managing Finance or you know
13:08 they they may use data but that's not
13:10 their main number one job uh in the day
13:12 that's ours right so how do we
13:13 communicate with those
13:16 folks so and and there conceptual data
13:18 modeling M I think in data modeling and
13:19 hopefully we get over this soon I'm
13:21 guilty of it myself we almost feel like
13:23 we have to defend ourselves for why
13:25 we're doing this but it is at the core
13:28 of so many things so a big driver in the
13:30 growth of data and we are seeing growth
13:31 there things like business intelligence
13:33 and analytics and and here's an example
13:35 of I think why a conceptual model is
13:36 important so say you know the
13:38 traditional business user they in the
13:39 corner saying can you show me all the
13:41 customers by region and I think we all
13:43 many of us in the call have understand
13:45 it atypical way we get this at least
13:46 historically for data warehousing you
13:48 kind of have your Source systems and a
13:50 data model helps there with getting kind
13:51 of the data structures and how it's
13:54 stored and structured um you might put
13:55 that in the dimensional dat warehouse
13:57 and kind of get your starc run in fact
13:59 we talked about that last month
14:02 um and that's it could be busy but it's
14:04 sort of on the on the technical side so
14:06 we we decided in this hypothetical
14:07 situation and we're going to start with
14:09 a conceptual data model because we saw
14:11 Don's dat diversity webinar and we are
14:12 all gung-ho about conceptual data
14:15 modeling so we started off with just a
14:18 very very simple model with the basics
14:20 what do we mean by customer right and we
14:22 spent a lot of time on this definition
14:24 because if anyone who builds definitions
14:25 knows that there's an art to this and a
14:27 bit of a science as well it's not just a
14:30 customers a c customer or custom you
14:31 know you know the very simple
14:34 definitions a product ID is an ID of a
14:36 product we've all seen those that aren't
14:38 very helpful so we one thought it's a a
14:39 person or an organization you know we
14:42 could do B2B or b2c they've purchased
14:43 one of at least one of our products
14:45 could be more they have an active
14:46 account now that that could be important
14:48 maybe they're a product but they they've
14:49 dropped maintenance or you know
14:51 something simple like that so we spent a
14:53 lot of time on this and we were wearing
14:54 it well you know are people going to
14:56 think this is a little too simple um is
14:59 this too academic um
15:01 and they did unfortunately this team
15:03 didn't really understand you know they
15:04 thought we were okay you academics up
15:06 there in your Ivory Tower you can go
15:08 debate the definition of customer duh
15:10 that's so obvious we're going to go do
15:12 the hard stuff we're going to you know
15:13 build all that warehouse stuff we talked
15:15 about in the previous slide they reverse
15:16 engineered and created physical data
15:18 models for each systems that was great
15:20 and they created the ETL scripts and
15:22 migrated to the warehouse and then you
15:23 know one of the key things I always
15:27 stress um in all my whenever I can is um
15:29 you the business value of this we don't
15:31 do data just for philosophical reasons
15:33 it's generally for some business value
15:36 um so often it's hard to kind of get Roi
15:38 from this so we were actually saying you
15:40 know that bottom bullet if we can
15:42 actually send out a welcome email and
15:44 give people a coupon and say we can
15:45 actually show Roi that we built this
15:47 reporting warehouse and these are the
15:49 actual results of all the people that
15:50 purchased something as a result so I we
15:53 were pretty clever for doing that um the
15:54 impementation went perfectly the scripts
15:56 were fine you know we had done all the
15:58 physical stuff great because we did have
16:01 the model behind that well until we
16:03 showed it to the sponsor uh and of
16:05 course you know the business folks they
16:06 know their customers more than anything
16:08 else you know just gut feel if nothing
16:09 else and she said you know we can't have
16:12 2,000 customers in this I noce gut feel
16:14 we have about 400 and Jones TI they were
16:16 actually evaluating our product for a
16:19 10% Global discount you just G 50% off
16:20 you know some sales rep is going to find
16:23 you uh your house and break your
16:25 kneecaps because you just really ruined
16:27 their commission um the main thing is
16:28 you spent all this money in it to build
16:31 the port and the data was bad um and
16:32 that's always me in the front the one
16:33 getting the Heat and the guy in the back
16:37 like nope not me not me anyway um what
16:38 did go wrong right we did All the Right
16:41 Stuff technically by the textbook well
16:44 typical what do you mean by customer um
16:45 and before anyone laughs or think this
16:47 is too simple I have worked for
16:49 companies that have made this mistake
16:50 and I don't want to call out a name
16:53 because I'm sure it's it happens often
16:54 um unfortunately these kind of basic
16:57 definitions what was happening is we had
16:58 our quote customer database and our
17:01 quote customer database used by sales
17:03 those were actually prospects right
17:05 that's an easy thing I mean most sales
17:06 people say I'm going to go visit a
17:08 customer today well often that's
17:10 actually technically a prospect you mean
17:12 who who corrects them right but you know
17:13 if we're going to put in a database
17:14 those are actually different things so
17:15 what happened this is actually a
17:17 business result we sent a discount
17:19 coupon to 1,600 of the wrong people we
17:20 gave upper management Report with the
17:22 wrong finger figure and now we actually
17:24 have to go back and fix it right so
17:25 what's that saying if you don't have
17:27 time to do it right you have time to do
17:29 it again right so a lot of problems just
17:31 kind caused by something as simple as a
17:34 bad definition so this time we started
17:36 again with a conceptual data model that
17:37 with our beautiful definition that's
17:39 there so a prospect is a person or
17:41 organization who doesn't own a pro
17:43 product is looking at it and a customer
17:45 is someone who does and has an active
17:48 account so again super simple business
17:50 definition but very very important
17:52 business results so again hopefully this
17:53 very simple business
17:55 definition uh kind of showed that this
17:57 does have an actual business impact and
17:59 the other thing I talked about this a
18:01 lot in the Enterprise architecture
18:02 webinar I gave a couple months ago this
18:04 doesn't have to take forever I mean it
18:06 can I mean some of these things maybe we
18:07 don't agree with the definition of the
18:09 customer and it takes some iterations
18:10 but sometimes it's an afternoon
18:12 whiteboarding and just taking that first
18:15 step and starting um go a long way and
18:17 yes the metadata repository behind it
18:19 and published models is all great but I
18:21 I wouldn't be afraid of even starting um
18:23 because sometimes you can dis fles out
18:25 some really simple stuff by a simple
18:27 white board in an afternoon or hour with
18:31 some folks um so uh the importance of
18:32 business definitions you've probably
18:34 seen these cartoons in other forums
18:36 before because I'm hey if you have data
18:37 modling cartoons use them where else can
18:40 you see data oning cartoons um and this
18:41 probably isn't funny unless maybe it's
18:43 not funny at all um but unless you're in
18:45 the business right we're all all done
18:46 with acceptance testing and everything
18:47 looks great and this new marketing
18:49 application this little question what do
18:51 you mean by customer well as we just
18:53 showed that it's really hard to fix
18:55 after the fact you built the whole
18:57 system and your basic definitions are
18:58 wrong right so get the basic
19:00 requirements before you start and that's
19:03 where a conceptual data model and
19:06 help another data modeling cartoon um so
19:08 the other thing and I may differ with
19:11 some others in the industry use the
19:14 language of your audience in the model
19:17 um so a couple things one is display a
19:18 way that's intuitive to so some folks
19:21 say a business person can't understand a
19:22 model you know two type well you should
19:24 build it in a way they do we use
19:25 PowerPoint all the time right we think
19:28 in pictures we tend to um understand
19:30 thing so a lot of I've seen
19:32 presentations from business users they
19:33 didn't call it a data model but it sort
19:35 of looked like one you know they might
19:36 have even boxes with things like product
19:38 and customer alliance between them
19:39 they're just trying to sus out their
19:41 things themselves um and it becomes very
19:43 much like a conceptual data model you
19:45 know put it in the PowerPoint do the
19:46 model and put it in PowerPoint if that's
19:49 how folks want to see it um use business
19:52 terms and avoid excess detail which I
19:53 think it's a beauty if you think that
19:56 pyramid again keeping it simple um and
19:59 to totally show show my ultimate nerdery
20:02 this is one of my favorite quotes a shoe
20:05 ofation so if you're not familiar with
20:07 what that means it basically means avoid
20:09 using big words to over complicate a
20:12 simple term which is sort of I don't
20:14 know I live in Boulder Colorado where
20:16 there are many other nerds I guess
20:17 because this is actually a bumper
20:19 sticker that people have um and in
20:21 writing circles this is of a joke that
20:23 people use you know if you can use 10
20:25 words but you could use two use two you
20:26 don't use a big word just to make
20:30 yourself sound smart and I think uh we
20:32 do that sometimes in technology if we
20:33 just use a lot of big words and show
20:35 people how smart we are in technology
20:38 won't they think we're great no actually
20:40 I think the opposite I think we scare a
20:41 lot of business people they don't want
20:43 to talk to the tech it either we're
20:44 talking down to them or we're talking
20:47 that geek stuff so you know asso obus
20:50 keep it simple stupid I guess is the the
20:53 quick version of that um and Talking
20:55 your business's language so this raises
20:57 the Eternal question that I know keeps
20:59 you up at night um
21:01 can and should a business person learn a
21:03 data modeling notation so can I think I
21:04 already talked to that yeah of course
21:06 they can it should not be hard this is
21:09 not brain surgery um I think the beauty
21:10 of a model is that it's simple the
21:12 concepts that we're describing can be
21:15 hard uh should should they I would say
21:16 we should build the models in a way that
21:19 they don't have to um but they can and
21:22 I'll go through a quick you know 101 of
21:24 modeling for those folks who might be
21:25 business people on the call who might be
21:28 familiar with modeling Andor a nice way
21:29 for you to describe modeling to your business
21:31 business
21:35 stakeholder um so here it is so again if
21:36 you've never seen modeling this is a
21:38 little primer for you but I think more
21:39 importantly for the technical people in
21:42 the call you should be able to explain
21:44 your model uh to a business person in
21:45 about 5 to 10 minutes quickly with
21:47 slides like this and I think they should
21:49 be able to get it people are smart um
21:51 and the model should be simple so an
21:53 entity those are your nouns of the
21:54 business right the who what where why
21:56 when the who might be a salesperson what
21:58 might be a product you know know how
22:00 we're invoicing people through an
22:02 invoice uh there's the dma dictionary
22:04 definition of it but again I wouldn't
22:06 start there I just say this the nouns
22:07 right you might even start looking
22:10 through requirements documents to or I
22:11 mean business documents when you build
22:13 your model and I often do this just draw
22:15 boxes around the nouns right those are
22:16 often your business entities I keep
22:18 hearing about products and this well we
22:21 must it's probably an entity right so
22:22 nouns are the
22:25 entities um attributes or or ways to
22:28 further describe that entity so again
22:29 employee last name first name higher
22:33 date again this is fairly basic I think
22:34 most business people would understand
22:36 this is the descriptions the adjectives
22:38 about the thing um and even the model
22:39 there at the bottom that's pretty
22:41 straightforward we don't have to get
22:42 into the fact that that thing on top is
22:45 a primary key but if you did just call
22:47 us a unique identifier for this thing
22:49 you know people can understand that um
22:51 and then the attributes are around it so
22:53 that should be fairly straightforward to
22:57 people um the next one are the verbs of
22:58 the organization
23:00 um and if any of you uh have heard my
23:02 full story which I'll shorten here um I
23:04 was probably one of the few I don't even
23:05 know if they teach it in school anymore
23:07 but diagramming a sentence I think I was
23:09 six years old and they did the
23:11 diagramming sentence which kind of um
23:13 you you underline the verb and you you
23:16 Circle the nouns and you put a downward
23:18 line to a prepositional phrase very much
23:20 like data modeling actually so I knew I
23:21 was a data modeler back then I think it
23:23 was the only kid in class that really
23:25 got into it I still diagram sentences
23:28 time to time um in a way it's like often
23:30 start with drawing a box around the
23:32 nouns one at the bottom a department
23:34 employee draw a line under the verbs
23:36 that's really your your relationship
23:38 line right a department can contain more
23:40 than one employee again you can take a
23:42 lot of these business rules that are
23:43 part of your organization that might
23:46 exist in some documents and really kind
23:47 of easily that way start turning it into
23:49 a data model Department can contain an
23:51 employee a customer can have more than
23:54 one account those are you know uh those
23:56 are the relationships the sentences of
23:57 your business and some tools are you
23:59 going to write scripts around the tools
24:00 that can take your data model and
24:03 generate these type of sentences so we
24:05 just did a workshop in our practice
24:06 actually I wasn't there my team in the
24:10 UK did this but um where we took the
24:12 model and created these business roles
24:13 when we had the meeting with the
24:14 business stakeholders we read them the
24:16 sentences we didn't necessarily well we
24:18 also showed them on model but to kind of
24:19 check the model is you know does the
24:21 sentence make sense to you uh no a
24:23 customer can only have one account the
24:24 customer and account those are almost
24:26 the same thing in fact we could call an
24:28 account customer so wrong you might have
24:32 more than one um you know wallet we call
24:33 them wallets what you call an account
24:35 you know all those kind of thing they
24:37 get flushed out by sometimes just these simple
24:39 simple
24:41 sentences um and then cardinality that's
24:43 a great big word um that we can make go
24:46 that issue OB we can make a really
24:47 simple thing sound really techy and
24:50 scary um all that is is the how many
24:51 right and I like this little picture
24:54 it's if you if you're using the IE um
24:56 notation instead of look if you look at
24:58 that on the employees that the one to
25:00 many one is really if you think of a kid
25:02 how many they hold up their hand right
25:04 one is one finger the other one you
25:05 could turn almost looks like a hand
25:07 several fingers I actually and I am fun
25:09 at parties let me tell you this is the
25:11 type of stuff I talk about um but when
25:12 we did the data modeling for the
25:15 business book uh again we were curious
25:16 there's a lot of different notations
25:19 some people like I def some people like
25:20 iie you know there's a lot of different
25:22 ones I actually created a simple
25:25 business rule and put it in a data model
25:26 and ran up by a bunch of my friends
25:29 again I am fun um and folks that one was
25:32 a painter one was a sculptor one was a
25:34 musician One Was An Architect that built
25:36 houses there was a finance person one
25:37 was an engineer none of them were in
25:39 data per se and I kind of said what do
25:42 you think this means most of them looked
25:44 at the one that was the the notation
25:45 here and could kind of get it well
25:46 there's one thing and there's kind of
25:47 many things and it looks like a
25:48 department can have more than one
25:50 employee I me a lot of people got it
25:52 without my even explaining anything they
25:54 just by the notation so I'm a big fan of
25:56 this one but again let's not be too
25:57 academic I just find that when I'm
26:00 talking to business users this one seems
26:03 to kind of make sense for folks so I
26:04 tend to like information engineering
26:05 which is crows feet a lot of people call
26:09 it um or children's fingers what I call
26:11 that um it kind of looks like a crow's
26:12 foot as well but that doesn't show the
26:15 the menu so anyway that is it I mean how
26:17 long did that take right here's another
26:19 one super types and subtypes can sounds
26:21 so complicated but it really isn't it's
26:24 either or or and right so here maybe
26:26 just draw an example a vehicle if it's
26:29 an exclusive or is it X exclusive again
26:33 in this notation um can be a car or a
26:35 truck it cannot be both we can argue
26:36 there some cars that are some but in
26:38 this particular business there are cars
26:39 and there are trucks and they're not the
26:41 same thing you can't be a car and a
26:42 Truck at the same time you're one or the
26:44 other or you could have an inclusive
26:47 subtype where I'm a person and I can be
26:50 a customer and I can be employee of that
26:51 company and that might be kind of a
26:53 thing to talk about oh wow if our
26:54 employees are buying the product they
26:56 get a discount do we not let them buy
26:58 the product we you know could be a
27:00 conversation right there um and just as
27:01 a joke I I was a spent a time in
27:03 marketing at a data modeling company and
27:05 my boss at the time was not a data
27:06 modeler she had grown up in marketing
27:09 pure business person um wanted to know a
27:11 little bit of modeling and I would I put
27:13 this in a presentation and at first
27:14 she's like what's that X and I just
27:17 explained she like oh um it made a lot
27:19 of sense it just was a way to explain we
27:21 were talking about campaigns and we were
27:23 segmenting customers and this is my way
27:24 of saying we don't want prospects we do
27:27 want customers that kind of thing um and
27:28 she it and like she kind of joked oh
27:29 that's an exclusive sub type
27:31 relationship and that was sort of a joke
27:33 the rest of the uh time I worked for her
27:34 she was a marketing person she loved to
27:35 talk about exclusive sub type
27:36 relationships that's probably rare you
27:39 should you should not use that type of
27:40 terminology um when you're talking to
27:43 business um but the point is this is a
27:46 particularly easy one to understand and
27:48 you know I've heard folk again I might
27:50 go outside the lines when we say what's
27:51 purely a conceptual model there's been
27:53 an argument of you know are should super
27:55 typ sometimes be in the conceptual model
27:57 I think so because I think it makes a
27:59 lot of sense you're you're starting
28:02 to flush e can talk business rules like
28:04 can a car and a Truck be the same thing
28:05 or a car manufacturer what about these
28:07 hybrid vehicles could there be a third
28:09 thing you these are very important
28:11 Concepts and it's pretty easy to
28:12 understand so yes I would say I'll put
28:13 myself out in the limb that yes you
28:16 could put um a super type sub type in a
28:18 conceptual data
28:20 model the other thing that might cause
28:22 some disagreement with me but I'll stick
28:25 to it um is it do we use the business
28:27 terminology or you know a common thing
28:29 in the industry is this idea of say a
28:31 party right the beauty of the party
28:34 model that sounds like it's a good time
28:35 is you know you you could have a there's
28:37 there certain things about a customer an
28:41 employee or or a client and a customer
28:42 that are similar can we roll them all up
28:44 into this concept of a party that's a
28:46 reusable thing that's a great idea to
28:47 use but I would say when you're talking
28:50 to the business person unless you know I
28:51 might be legal and there's a party in a
28:54 dispute and they use that term party for
28:56 their customers that means something
28:58 then I would use it but say had a party
28:59 associated with an entity what the heck
29:02 does that mean it could be a legal party
29:04 is associated with a a legal entity
29:06 which is the company unless it means
29:08 that it's very very vague it's a thing
29:11 relates to thing you could you could
29:12 summarize things so much that you just
29:13 said you know things relate to things
29:16 and everything's a thing and that gets
29:19 crazy right so I particularly prefer to
29:21 say things like a customer PR a product
29:23 or employee works for a department
29:25 that's the terms the business people are
29:27 using and maybe there's some redundant
29:29 see when you use the actual physical
29:30 model maybe you want to do that
29:31 differently but at this point we're
29:34 trying to get the difference between a a
29:35 you know business terminology so I will
29:37 ask the question I ask a lot of
29:38 questions when I do this is a customer
29:40 the same as a party is the customer the
29:42 same as a client is the customer the
29:43 same as the pro you know that that's the
29:46 type of stuff you're trying to flush out
29:47 um you could have two versions of
29:49 customer on a conceptual data model with
29:51 two different definitions and trying to
29:52 say are they different you know this is
29:54 might up to the endgame but you're
29:55 trying to understand okay maybe that's
29:57 Europe's definition of customer in North
29:58 America America uses something different
30:00 let's try to understand that more again
30:02 you're just trying to communicate when
30:03 you're talking about this conceptual
30:06 data model and keep the focus in the
30:08 business as I mentioned we can often get
30:10 academic I mean one of the reasons we
30:12 like or I like the other thing it is
30:14 that kind of logic we go through but
30:16 don't make it a logical exercise it
30:18 should be a business exercise so if
30:19 you're arguing let's think why we're
30:22 arguing is it a different entity then
30:24 yes we should really flush this out is
30:26 it different names for the same entity
30:28 Maybe yes that's something we should be
30:30 discussing could it be a super type
30:32 subtype relationship and how you resolve
30:34 that could be different ways maybe we
30:35 keep the different names maybe we try to
30:37 make it one name but again you should be
30:39 arguing where it's a business definition
30:43 not semantics or just theoretical we can
30:44 I I have to catch myself sometimes it's
30:46 just sort of fun to start to think these
30:47 through but at some point you're like
30:48 does this matter does this matter to the
30:50 business or we just being academic so as
30:53 a client a customer those the same thing
30:54 what's the difference between an
30:56 ingredient and raw material we had a
30:57 customer we give an example in the state
30:59 of Ming for the business book that there
31:00 was an argument about that right some
31:03 folks is raw sugar an ingredient into a
31:05 piece of candy or is it a raw material
31:07 shipped from Brazil that's actually cane
31:10 right so could be either one there's no
31:11 answer to that the answer is how the
31:14 business uses it um again I mentioned we
31:16 just did a modeling workshop with the
31:17 our team in the UK and there was
31:19 actually an argument whether water was a
31:22 liquid so I mean this was a uh
31:23 environment group that was doing
31:26 environmental testing so it made sense
31:27 but I think I joked them I said I hope
31:29 your business sponsor didn't walk by the
31:31 room when you're arguing with water is a
31:32 you know some of these things can seem
31:34 really academic so I think we should
31:36 just check ourselves and say are we
31:38 arguing this because it means something
31:40 to the business that we need to resolve
31:42 or does it really not matter if you call
31:43 a customer or client is the same thing
31:45 just pick one and move on you know I
31:47 think the move on is sometimes what we
31:48 have to remember when we're doing is it
31:51 should not draw out unless it needs to
31:52 it should be as quick as possible to do
31:54 these kind of
31:57 models definitions are important so
31:59 don't slack on them um I think this is a
32:02 lot of the part of of the model that can
32:03 be difficult and to think through and as
32:05 we showed in the example you can have a
32:07 lot of big business issues caused by
32:10 just something like a IL defined term um
32:12 you know what do you mean by customer
32:13 would beat that one to death you know
32:15 how are R defining household if you've
32:17 been doing that in your business there's
32:19 a lot of different ways is it family
32:20 members is it people that live with the
32:22 same address you see I like this stuff
32:24 because it's just fun to think through
32:26 all the different combinations or
32:27 something as simple as how calculating
32:30 total sales and again it doesn't mean
32:32 that you're you're the folks at the top
32:34 defining this for everyone lots of times
32:36 it's just showing it I have some
32:38 companies working it's a particular tool
32:39 I like in the market that will show all
32:41 of the different definitions and just
32:43 the why I'm using this definition of
32:45 total sales for this report and this one
32:47 for a different report and this why
32:49 because you the different definitions
32:50 and and as long as you're clear about
32:52 that that's fine it's not you're
32:54 dictating necessarily um how people
32:55 should do it it's just being clear about
32:57 it you know sometimes you need to dict
33:00 but not all the time um and then any
33:01 Italian speakers in the phone you'll get
33:04 my little joke about appy API means B
33:07 Italian I was that was kind of cute um
33:08 so anyway but you know if we're in the
33:10 financial industry what an equity
33:12 derivative what's one of my first big
33:14 metadata projects was just listing all
33:16 these financial terms and what they mean
33:17 for the Brokers and then we built the
33:19 bit the technical stuff behind it but
33:20 the hardest thing was getting the
33:22 definitions right for some of these
33:23 terms and making sure every was
33:25 calculated in the same way so don't
33:27 slack on your definition that's a huge
33:31 part of conceptual data modeling um some
33:33 tools that actually you can show it on
33:36 the model I'm a big fan of that um
33:37 because then I think you know I think
33:39 sometimes showing something like customer
33:40 customer
33:42 account that just okay yeah but when we
33:44 start to say you know customers a person
33:46 organization with an active account
33:48 how's that different from a client or a
33:51 client has an active brokerage account
33:52 uh no we can roll those into the same
33:53 thing you know again that's what the
33:55 business person would have to make that
33:57 decision but often until you can't
33:59 until you see them you know is a broker
34:01 different from a salesperson oh wait
34:02 yeah you're right I think clients are
34:04 different because those are the high net
34:05 worth individuals and a salesperson is
34:08 any account under $100,000 US you know
34:09 something like that but unless you see
34:12 the definitions it's often hard to see
34:14 why you have client customer on the same
34:16 model um so anyway just a tool I like to
34:18 use it could be anything you could
34:19 export these out into a spreadsheet I
34:21 would say start with a model but again
34:23 it's however your audience wants to
34:25 consume them because the the big reason
34:27 we're doing this is for that community
34:30 so again if a a sign out in the street
34:31 is going to get people to read this I
34:33 would I would do that but I don't think
34:35 you have to try that hard often I think
34:37 that like as the quote said earlier a
34:38 lot of business people are actually
34:40 relieved that someone's thinking about
34:43 this um as long as you keep it simple I
34:44 don't have a slide for this one but it
34:46 was a thing you know respect people
34:48 times make the I think a lot of business
34:51 people if they are none willing to look
34:52 at your model might have had bad
34:54 experience in the time in the past I
34:56 have been in meetings where there's a
35:00 logical and priz model with 200 entities
35:01 printed on the wall would take up two
35:03 walls and we say to the business person
35:05 we're just going to spend an entire day
35:07 going through all of this and if you
35:07 don't mind we're going through the
35:10 cardinality and the relationship you no
35:11 surprise they don't want to do that
35:15 again I I had one customer that he would
35:17 put five entities on a blotter you know
35:18 when when his sponsor came in it would
35:20 have it on her desk he' just say five
35:21 entities 5 minutes and you could just
35:23 look at these really quickly and you
35:25 just sort of do a little subject area
35:27 every morning just I I just
35:29 if I could have 15 minutes your time or
35:30 whatever often some of these questions
35:32 don't take a long time if you can do it
35:35 in small chunks and more importantly do
35:36 it something that makes sense to them
35:37 that's a business problem hey I know
35:39 you're doing this marketing campaign can
35:40 I just double check with you when you
35:42 say customers you mean existing
35:44 customers and Prospects or just existing
35:46 customers that wasn't hard um but it's
35:48 just asking the question or if you need
35:50 to model a small model like this I just
35:51 wanted to know what do you mean the
35:53 difference between customer and client I
35:54 I drew this out does this make sense to
35:56 you something like that sometimes that's
35:58 all you need it doesn't have to be a
35:59 whole day workshop with you know
36:01 thousands of entities on a wall because
36:03 I would hate that as well no actually I
36:07 would like that but I'm I'm weird um and
36:10 and again human metadata I always say
36:12 avoid that dreaded I just know I cringe
36:13 when I hear that we don't need to Define
36:16 customer I get that well you get that
36:17 but your definition may not be somebody
36:19 else's you know so a lot of this
36:21 metadata in a company is an employee's
36:23 head you know the guy in the picture
36:24 well part number oh that used to be
36:25 called a component number those are
36:27 really the same thing well he knew that
36:30 then anyone else putting that in the
36:32 model and making you know this tribal
36:34 knowledge Enterprise knowledge is a huge
36:36 part of the model So it's talking to
36:37 these people it's it's getting it in the
36:39 model having the different pieces of it
36:41 it might be showing this person a small
36:44 model kind of making sure excuse me that
36:46 this is kind of published out to the
36:49 organization and here's my little my my
36:51 counseling for the day I I think of this
36:52 wouldn't the world be a better place if
36:54 we all did conceptual data modeling
36:56 because it really helps communication
36:58 and wouldn't this help our daily life
37:01 too so one of the weird things we do in
37:03 the US for some of us uh I never have
37:06 had to do it but children have a summer
37:08 vacation and parents say what's this
37:10 great idea we've got this lovely large
37:11 country we're going to drive across in
37:13 the summer we're going to see all the
37:15 different sites and I think many a
37:18 divorce and many a you know brotherly
37:20 sisterly Feud have been caused by being
37:22 stuck in a car for weeks on end driving
37:24 across the United States um so wouldn't
37:26 it have been better if we all said let's
37:28 go in a family vacation what do you mean
37:30 by vacation right so maybe the father is
37:32 like you know I think this is a great
37:34 opportunity to learn new things and I
37:36 want to step in every state park in
37:37 every state and learn a new fact in each
37:40 one and mom goes you know I really have
37:42 been working really hard I just want to
37:44 read a book so you can you know go to
37:46 your stupid state park I'm going to sit
37:48 in the car and read a book and Jane has
37:51 been studying at University for a while
37:52 she goes you know Dad you can stay in
37:54 your stupid Park and Mom read your book
37:55 I'm going to go out and take a run
37:57 because I have been studying in for
37:58 weeks and I just want to get outside in
38:00 the state park I'm not going to go look
38:02 at the interactive exhibit dad that's so
38:05 stupid um and Bobby didn't want to go at
38:06 all he's like for me vacation staying
38:08 home being with my friends so could I
38:10 just skip this whole thing all together
38:12 and Don's like well as long as I have my
38:14 laptop um and I can be building data
38:15 models in the car then I'm good right
38:17 and then Ian the Brit he's like you
38:18 Americans I call that a holiday you
38:20 don't even know the right word right so
38:22 all of this conflict uh just from
38:23 something as simple as what should have
38:25 been fun the vacation so if you think of
38:28 it if we often find our terms right uh
38:30 let's let's go to a party on Friday what
38:32 do you mean by party um maybe you could
38:34 be annoying but I think you can see that
38:36 sometimes just getting basic definitions
38:39 of things before you start a project uh
38:41 can avoid issue so um feel free to dat
38:43 them Mar with your family if that is
38:46 going to help things hopefully an
38:48 illustrative example so the other part
38:50 we've been talking about a bit is that
38:51 we are visual creatures and the other
38:53 Beauty you I mentioned you can put some
38:55 of this in a spreadsheet that's fine um
38:57 but the beauty of a data model is that
38:59 it is graphical and we tend as humans
39:01 think in pictures in fact this is an
39:04 actual cave drawing found in southern
39:06 France no just kidding it was actually
39:10 an Ida notation that one uh bad joke U
39:11 but we do we had to draw we TRW the
39:13 thinking pictures we draw pictures so
39:14 that's one of the beauties of a
39:16 conceptual D model is that is very
39:19 graphical and pictoral so one of the
39:22 things I do uh in most of my Consulting
39:24 projects and it it's something that most
39:25 people if any of my clients are on the
39:27 phone they may get
39:28 because most people do the first time
39:30 until they look at it and then they're
39:31 like oh this makes a lot of sense so I
39:33 often just put literal pictures of
39:36 things in a model right and this has
39:38 been very helpful at several clients uh
39:39 so you know I'm talking about a
39:41 salesperson that sells a product Oh
39:43 product and a box no no no no our
39:46 products are all online games you know
39:47 we wouldn't actually have a box for that
39:49 anymore we used to sell them in stores
39:50 and now it's all online so there's no
39:52 okay well there's a you know thing right
39:55 there and salesperson support rep okay
39:57 the support rep on the phone but you
39:58 know sales people are actually on the
40:00 phone too we don't have any you know so
40:04 I I I've had several customers where it
40:06 you know things came out of the model
40:10 just uh from drawing it out this way um
40:11 the other thing that sort of makes it in
40:14 fact it it sort of makes a drawing of
40:16 what the business is so several of my
40:17 clients I have one I did some work for
40:19 an outdoor industry and it was Stefan
40:21 Krauss and he worked in St maritz and he
40:23 was a ski instructor and he purchased
40:24 products so we had the thing all around
40:26 Stefan there was also a Stefan cruss who
40:29 is an accountant in zerk um he had
40:30 certain characteristics in a certain
40:32 salary and because he was in the Europe
40:35 he was subject to gdpr you know all the
40:37 sort of stories around that customer
40:39 especially when you're trying to explain
40:41 you know the connections between data or
40:44 why it's important um I think it really
40:45 starts to click with a business person
40:47 the customer down the right was he was
40:50 Stefon no he was Martin stes he was the
40:51 high Networth individual who had
40:52 different accounts in different
40:55 countries and he had yacht insurance and
40:57 all these sort of things um and it's
40:58 funny during the product people start to
41:00 say well well Stefon would you actually
41:03 start to sort of relate to it but that's
41:06 the purpose right um the client I met
41:08 today is a healthcare provider and their
41:10 their complaint was the the picture we
41:13 had the guy looked too healthy like how
41:14 many hospitals actually people are
41:16 actually just very attractive with a
41:17 Band-Aid on right you don't actually
41:19 have sick people but the story was
41:21 around you the hospital visits and the
41:23 the different claims and all that so it
41:25 really is is kind of a nice one pager at
41:28 a pictoral level um I've had different
41:31 aha moments with different um business
41:32 people one was a l us describe our
41:34 customer generally they're younger or
41:36 they're older or so you could laugh that
41:37 that's the feedback they had or you
41:39 could say wow we got them engaged
41:41 they're stared to see their business in
41:43 this right or another business person
41:45 said oh so I get it the hard part of the
41:47 warehouse is the relationships between
41:49 those things right that we might have
41:51 been missing a relationship yes there's
41:52 no connection between a salesperson and
41:54 product we need to add that or you know
41:57 it sort of makes it very real um because
41:58 you're literally seeing your business
42:00 and the reason I started doing these is
42:03 a am a goofball and D it helped me when
42:05 I go into another customer new customer
42:06 new business May haven't worked with it
42:09 helps me tell the story of what is this
42:11 data doing why are we doing it how does
42:14 it flow um and I said well might as well
42:15 show it to other people and there's my
42:17 tip for you I think it really does work
42:19 because it it kind of makes it very
42:21 concrete um for the
42:25 business um because we do tell stories
42:27 not only tutorial but everything is a
42:29 story right we can't even sleep without
42:31 dreaming which is really stories right
42:33 and so I think that the impact of this
42:35 is what I'm trying to say is they no one
42:36 cares about your data model you might
42:38 have heard me say this before and it's
42:40 sort of sad um but they do care about
42:41 the results they don't not that they
42:43 don't care about data modeling um they
42:45 want to see the real world impact what's
42:47 the story why are we doing this what are
42:49 the results of that model and what we're
42:50 trying to say is that we can't like
42:52 sales to salesperson because there's no
42:53 relationship there or if we could think
42:55 of all the great things we could do and
42:57 I think sometimes we missed the so what
42:59 um when we do modeling we can stay too
43:02 much in the database level so um you
43:03 know I'm not saying you necessarily read
43:05 data models to your children before they
43:07 go to bed but but the the point is it's
43:10 not the model it's the message um and I
43:12 had to remind myself in fact many my
43:13 training classes I kind of say that I'm
43:16 sorry but nobody cares about your model
43:17 and I I was working with a customer and
43:19 we do we get into what we're doing it
43:20 could be anything it could be a thing
43:22 you're writing an article you're writing
43:24 it could be your own taxes you know
43:25 anything you're in the middle of
43:27 something you're you're you're in it and
43:29 and the customer you know wasn't
43:30 connecting that day I think they
43:31 probably had a big meeting and I had to
43:33 joke to myself Donna nobody cares what
43:34 your data model you know what was a
43:36 story you were trying to tell so you
43:38 know I tease myself too it's like these
43:40 are the things you have to remember it's
43:41 why are you showing this to somebody
43:42 what's the
43:45 scenario um another you know you might
43:46 have seen my guy in the lower left but
43:48 somebody encouraged me that they said
43:49 they actually liked it though you're
43:51 going to see him again um but it reminds
43:53 me this is me you saw my picture in the
43:55 beginning but inside I'm that little guy
43:57 in the lower left
43:59 and so my point of this slide is that
44:01 there are different personalities and
44:02 different goals in the organization
44:04 everybody's that guy in the middle
44:06 what's in it for me right but I think
44:07 there's also different personalities so
44:09 we as data Architects we probably went
44:11 into data architecture because we're
44:12 focused on things like architecture and
44:16 technology and often just we're sort of
44:18 hired to fire find problems right we
44:21 find find problems and we fix them and
44:22 that can often make us seem very
44:24 negative right so business people
44:26 they're very results or oriented the
44:29 salesperson almost by definition is just
44:31 genetically positive you know the
44:32 customer said no I'm then to ask again
44:33 you know and and it's all about
44:36 opportunity and business growth so there
44:38 sometimes where we clash and we can be
44:40 seen as really nerdy that might be
44:41 surprising so you might be wondering
44:42 this little with this little guy I'll
44:44 explain it so we're often seen that kind
44:46 of that weirdo in the side of the street
44:47 you know holding up the sign going the
44:49 world is going to end if their model
44:51 isn't their normal form you we might be
44:53 right maybe the world will end if our
44:55 model isn't in third normal form but the
44:57 business person doesn't care you why
44:58 should I care that it's in third normal
45:01 form tell me the what so what so that's
45:03 my recommendation on the right can we be
45:05 the same person on the left
45:08 inside but but but put on the hat we all
45:10 wear different hats of more of a data
45:12 advisor less architect and more advisor
45:14 so think of the opportunities less hey
45:15 guys we can't do this with a modeler
45:17 this is broken or this is really hard
45:19 hey if we could Link customer day with
45:21 product think of the stuff we could do
45:23 um so what are we doing different on the
45:25 right well we're talking the business
45:26 person's language
45:28 why what do they want to do they want to
45:29 get more money they want to increase
45:31 sales so we could do this we're being
45:32 we're thinking of opportunities not
45:34 problems there could also be a product
45:36 problem we can't link customer with
45:37 product um and we're trying to get
45:39 funding to do that but we're saying it
45:40 in a positive way think all the great
45:42 stuff we could do if um because that's
45:44 often how really successful business
45:46 people live and work that's why they're
45:48 successful always that next same thing
45:49 and we can get annoyed by them because
45:51 sometimes they forget the details and
45:53 that's what we clean up after if we're
45:55 talking to them that's really what
45:56 they're thinking of and that's really
45:58 what I think is fun about data
46:00 management why I'm still in the business
46:02 um because especially with data now
46:03 there are so many cool things we could
46:05 do um so putting on mat hat especially
46:06 when you're speaking with the business
46:08 executive that's where you become data
46:10 adviser unless data architect who the
46:14 weirdo with with the sign in birken STS
46:16 which we all are part of probably all of
46:18 those right so just think of your
46:21 audience I guess is the message there um
46:23 and and we do the same thing right as I
46:25 mentioned before do we really we're all
46:27 into our model do we care about the
46:29 details of other people's jobs so think
46:30 of an accountant right so we talk to the
46:31 accountant we knock on his door he's
46:33 like you know we recently switched to a
46:35 Cru based accounting based on cash based
46:36 accounting because and you're like I
46:38 really just want my paycheck do I need
46:40 to know all the details of the
46:42 accounting system I I just want my
46:44 paycheck so again business people are
46:46 the same thing with a data model well
46:48 why would I care I just want the data
46:50 for this report or I just want to make
46:52 more money through sales how can I do
46:54 that how can data help me so you know
46:56 it's not that folks are bad people that
46:58 people don't like data I think actually
46:59 a lot of people are interested you just
47:01 really don't care about the details of
47:02 what we do because we do the same thing
47:05 I give any job you know I have a an
47:06 electrician come to fix my house I
47:08 really just want the light switch to go
47:09 on I don't know the detail you know
47:10 sometimes you are you kind of want to
47:11 know what they're doing but in general
47:13 you just want the results you don't care
47:14 about the
47:17 details so um again if you've been in
47:18 any of my workshops we often kind of
47:22 have this as one of the the uh exercises
47:24 and I think you know driving home think
47:27 of this or if you're at work or wherever
47:29 dinner think of this um how would you
47:31 describe your project to the CEO in two
47:33 minutes and this actually happened to be
47:34 once in the US we always said what
47:35 what's the elevator pitch right you're
47:38 riding up the elevator with the CEO and
47:40 they ask you what you're working on or
47:41 you have a sales Prospect in the
47:43 elevator what would you say to them in
47:45 two minutes or how fast your elevator is
47:46 to sell them so what you probably don't
47:48 want to do in the left is you know I'm
47:49 working on a project to rationalize
47:52 metata across data you've lost them
47:54 right that's really not very interesting
47:55 to other folks but if you can put it
47:57 their context well I'm working on that
47:59 big uh online marketing campaign you're
48:00 doing we're going to get a better
48:01 customer list so that you can have
48:03 better content targeting or you know
48:05 whatever it is you're working on think
48:08 of that the context targeting you know
48:09 my story is this actually happened to me
48:11 I was working at a custom one of my
48:12 first programming jobs way back in the
48:14 day when I was a software engineer um
48:16 and I was riding up the elevator with
48:18 the CEO and he did say almost literally
48:19 just this so what are you working on and
48:21 I told him about the program he said but
48:24 what does that program do and why and I
48:26 couldn't answer it so it was partly my
48:28 manager's fault for never filling Us in
48:30 with a big picture but it was partly my
48:31 fault for not thinking of the big
48:32 picture I was so focused on getting that
48:34 application to run I knew what my little
48:36 piece did but I really couldn't tell the
48:37 big story so that was embarrassing
48:39 enough that that never happened again um
48:41 and hopefully you think of it before
48:43 that may ever happen to you um or you
48:44 know in anything it's not just riding up
48:45 the elevator could be anything when
48:47 before you give your presentation before
48:49 you do anything what's the so what and
48:52 why do you care and why would people
48:54 care um so that that's basically it when
48:56 we're thinking of the you know a lot
48:57 more we could to talk about but again in
48:59 the spirit of keeping it simple um with
49:00 conceptual models we're focusing on the
49:02 business business stakeholders business
49:05 rules so focus on what is interesting to
49:06 them have your presentation suit the
49:09 audience keep it simple and think of the
49:10 story right well it's your elevator
49:13 pitch um and don't make it overly
49:15 complicated so before we over up the
49:17 questions just a little bit if you did
49:18 have a question and it doesn't get
49:20 answered today or you just want to say
49:22 hi um I'm here's my email I'm also on
49:24 Twitter as well as our company and my
49:26 personal website
49:28 um a little bit about my company global
49:30 data strategy we do this for a living so
49:33 if you want help let me know um there is
49:35 a metadata management course online in
49:37 addition if you're at edw in Atlanta we
49:38 all do the whole half day metadata
49:40 session um but if you just want the the
49:43 Nuggets online and in your pajamas you
49:46 can watch this um and just if you can
49:47 join us for any of the others there's a
49:49 whole lineup for the rest of the year if
49:52 you are interested so without further
49:54 Ado Shannon we can open it up for questions
49:56 questions
49:57 Donna thank you so much for another
49:59 fantastic presentation we had a lot of
50:01 great questions coming in to answer the
50:03 most commonly asked question just a
50:04 reminder I will be sending out a
50:06 follow-up email by end of day Monday
50:08 with links to the slides the recording
50:10 and anything else requested throughout
50:12 the webinar we did have a request for
50:15 the um the report the metadata uh report
50:17 that you were showing earlier so I'll
50:20 get that out to everybody as well so
50:22 just diving in here Donna to the to the
50:23 questions you know this C this question
50:26 came on um you may have addressed it a
50:28 little bit but um have you seen
50:31 conceptual data models done for the bi
50:33 later yes
50:37 and in fact I had a slide I took out um
50:39 and and that is something I've often
50:42 heard why isn't um bi your star schema
50:46 isn't that just a physical thing um and
50:47 I had a slide that actually looked like
50:48 a star and and I think when you're
50:50 thinking think of a star scheme I'm
50:53 reporting on sales by region by customer
50:55 all those things you're reporting by I
50:57 mean I often just do a very high level
50:59 model that is the star schema and make
51:01 sure do I understand just what those
51:03 Dimensions mean do I understand how
51:06 we're summarizing so I think yes and I'm
51:09 not sure why that is so often fought
51:10 with me that well no no no it's a of
51:12 course it's a physical you're building a
51:13 warehouse but if you don't get I would
51:15 think even more importantly with bi if
51:17 you don't get those core definitions
51:20 then your reports aren't right so I and
51:21 and try to just understand with the
51:23 business what are we reporting by and I
51:25 have just kind of a very simple you know
51:26 might be
51:27 you know the
51:29 the entity in the middle and some Stars
51:31 around it and that kind of makes sense
51:32 this is the main thing we're reporting
51:34 on this is the slices of the dimensions
51:35 we're reporting by so yeah I think it
51:38 makes a lot of sense for
51:42 business so uh as I know Dr Peter Chen
51:43 the inventor of er model diagram
51:45 considers ER model as conceptual and
51:47 wished it to be used by the business
51:49 people from the very beginning but with
51:51 very um what has been changed in since
51:54 Peter's time and approach
51:56 approach
51:57 um well I mean I think I would agree
51:59 with them that you know there are
52:01 business people that can understand a
52:04 pure data model um I think well I'm just
52:06 seeing a trend in the business in
52:08 general I mean I had to kind of switch
52:10 my thinking um you know often has been
52:12 very prescriptive you know you should of
52:13 designed and then you implement and you
52:16 build and I almost think the difference
52:18 between encyclopedia and Wikipedia you
52:19 know this whole idea of of
52:22 crowdsourcing um so I guess I've just
52:24 been more creative in my Approach and
52:26 and so you some of these things like
52:28 pictures I think you know I I've sort of
52:30 seen even just metadata repositories
52:31 where there's comments and threads and
52:33 almost like a slack channel from
52:36 metadata um so I think I think some of
52:38 the Core Concepts are the same um he's
52:40 done a lot for the industry but I think
52:42 and I I've can get my brain around it
52:43 often and say you know this is a good
52:46 thing it's not as structured might be
52:48 more agile um but I think so some of
52:50 these creative maybe it is a slack
52:51 channel for metadata maybe it is a
52:54 picture that just gets the idea across
52:56 and maybe you put in the model later or
52:58 maybe it's iterations of the model um I
53:01 think the smaller bite-sized chunks in
53:03 the add world with small amounts of time
53:05 happen so you know I think the Core
53:06 Concepts are the same I think we've
53:07 gotten a little more creative on the
53:09 front end for some of this when we talk
53:12 about the conceptual model my two cents
53:15 on that one sure so one of the
53:16 challenges I find in building data
53:18 models is showing the relationships in a
53:20 way that is simple and palatable for
53:22 business users do you have any opinions
53:24 or recommendations on software options
53:27 for develop data models
53:29 models
53:32 um the relationships I have two ways I
53:34 show those let's see if I can go back to
53:39 my screen um one answer because I think
53:41 the relationship well one answer as I'm
53:43 looking for my slide is I wouldn't I
53:45 show only the relationship that makes
53:46 sense I mean if you're doing a bigger
53:48 model sometimes it is those relationship
53:51 lines that get really gnarly the LA of a
53:54 better word very quickly um so that's
53:55 one thing and this many of the data
53:57 modeling tools in the market I never say
53:59 the name of any tool so stop asking me
54:01 if that people did not that you did I
54:02 often get that question I will not
54:05 answer that one um you can hide and show
54:07 on the same model different relationship
54:12 lines um but if we think of the one that
54:14 I had that was just the oh gosh I'm
54:16 losing my mind it's late in the day uh
54:19 just this kind of one that the the story
54:21 kind of one I often will do this in a
54:23 PowerPoint and and just this if I'm
54:26 really just breaking Down The Core
54:28 Concepts um I doing PowerPoint just so
54:30 PowerPoint line and this almost looks
54:32 more like a data flow and it's not
54:34 typically your one to any relationships
54:35 but it depends what you're trying to
54:36 show if you're just showing basic these
54:38 things relate to these things and the
54:41 connections are hard I show that if I'm
54:43 showing relationships I do it more in
54:45 the way that we showed with the IE um
54:47 and I will show crow's feet uh because I
54:49 think that does make a lot of sense I
54:51 think very quickly even on this one what
54:54 I like about that is that a customer can
54:55 have one or more employees and he starts
54:57 from the cad cardinality and I often
54:59 hide the detail you'll see here in this
55:01 modeling tool it's just just the name
55:03 and you know it kind of simplifies it so
55:05 I'm a fan of crow feet notation and most
55:07 of the data modeling tools have that so
55:08 it's not necessarily always the tool
55:10 it's either PowerPoint if you're doing
55:12 really simply or in the modeling tool
55:13 use the notation but just simplify it
55:15 don't show all the attributes and all
55:17 the relationships just a few of them
55:21 that makes sense sub sure yeah so um
55:23 when building an Enterprise conceptual
55:25 data model um should you capture all
55:27 attributes key attributes or no
55:30 attributes oh good question who answered
55:32 that one so that was that's another I
55:33 would say here's the great consultant
55:35 answer it depends no but I will say what
55:37 it depends on I think if you're just
55:39 trying to get this main I'm trying to
55:42 get Department of more than one
55:45 employee um then yeah I think sometimes
55:49 hiding that detail is is good um for
55:52 example sometimes if the I like to show
55:54 attributes if they're not overwhelming
55:55 probably not every single after for
55:58 example I'm talking about customers um
56:00 and I show gender code and people say oh
56:02 no we only sell the corporations that
56:04 wouldn't have a gender so I think if the
56:08 attributes show um help clarify the
56:10 meaning I definitely show them well what
56:11 what's the department well we have a
56:13 department head you know these example
56:14 might not be good because it might be
56:16 but were it not obvious U or here with
56:18 the example I had with a car or a truck
56:20 well what are the attributes of truck
56:21 that would be different from the
56:23 attributes of car and that might help
56:25 answer if they're different so if it's
56:27 helpful I add them I probably would say
56:28 you don't unless there's only a small
56:30 amount of attributes show all of them I
56:31 would say probably never show because
56:32 that's when you're starting getting into
56:34 logical and it gets overwhelming shouts
56:37 to either none or that small subset that
56:38 kind of adds Clarity to your discussion
56:40 of these are the things that seem to be
56:41 different or this is what makes up a
56:43 customer that's different than a client
56:46 or something like that so hope that
56:49 help indeed so you know how do we
56:52 incorporate this data modeling into the AA
56:53 AA
56:56 methodology I think it is perfect for
56:59 agile methodology um the partly when
57:01 when you're doing a um model writing it
57:05 is a small subset and and often when you
57:07 think of agile or a Sprint or a
57:09 requirements phase it's getting that
57:10 requirement from the business and
57:12 turning them into stories so this would
57:14 be at the very beginning just making
57:17 sure we're getting the right um right
57:19 requirements is a huge part of it and
57:20 the beauty of these Concepts they should
57:22 be iterative you should change them as I
57:24 I mentioned before but arguing arguing
57:26 is that thing that's exactly the point
57:28 the point is communication so it should
57:29 change around and at the beginning of a
57:31 Sprint or during a Sprint you can use
57:33 this to help kind of vet out the data
57:35 aspect of it and it should change and it
57:36 should generate discussion and kind of
57:38 turn into the stories it's the context
57:39 ban the stories you're you're doing in
57:42 your Sprint so I think it's perfect for
57:44 because it's so quick it isn't like you
57:46 have to develop a whole physical or
57:48 logical model you can just start to do
57:49 some of the conceptual stuff to Z out
57:51 just the pieces that make sense or get
57:52 clarity I almost always draw a model
57:54 when I start have questions do you mean
57:57 X or Y or Z and often that model can
58:00 help be Clarity generate Clarity is a lot
58:05 faster all right I think we have time to
58:07 sneak in one more question here uh do
58:09 you have recommendations considerations
58:11 for the taxonomy standards selected for
58:16 model
58:20 um no I mean I think I mean taxonomies
58:21 can mean different things for people
58:22 some people like it kind of the super
58:26 type subtypes um I would say whatever
58:28 makes sense often taxonomies are good
58:29 outside of a model sometimes kind of
58:31 just showing a hierarchical approach and
58:34 kind of um a list a structured list help
58:36 so I think whatever kind of makes sense
58:38 for the business I kind of sometimes
58:40 like this approach kind of a hierarchy
58:45 kind of taxonomy um I think I think also
58:46 sometimes everyone me something
58:48 different about some tonomy which I'm
58:50 being vague in my answer um the other
58:51 thing that often comes up is if I have
58:55 different meanings of a a thing um a
58:56 customer could be a client you know I
58:59 often will just show kind of dotted line
59:00 relationships to those rather than try
59:03 to I think oftentimes people try to
59:04 force everything to be the same or it
59:06 might just be a taxonomy or a hierarchy
59:09 so I think sometimes the uh conceptual
59:11 model can help flush that
59:14 out well Donna thank you so much for
59:16 another fantastic presentation and
59:17 thanks to our attendees for being so
59:19 engaged in everything we do we just love
59:21 the questions that coming in but
59:23 unfortunately that's all we have time
59:25 for today just a reminder I will send a
59:28 follow-up email by end of day Monday
59:29 with links to the slides the recording
59:31 of the session and also include the
59:34 research report that we uh worked with
59:36 on with Donna that she talked about on
59:38 metadata and thanks to idear for
59:39 sponsoring today's webinar and helping
59:41 us make it all happen we appreciate it
59:44 and I hope everyone has a fantastic day
