0:09 that'll work for you yeah thank you very
0:12 much uh hey good morning everyone and um
0:15 for those of you that don't know gangas
0:17 uh they're up there on the right on the
0:21 top uh and uh gangas is actually a uh
0:24 decision intelligence organization so we
0:26 work with clients to help them gather an
0:28 gather data we work with clients to help
0:31 them do uh data sciences and analytics
0:33 on that data and then we work with
0:36 clients to help them um curate their
0:39 insights and uh help them to be more
0:41 persistent uh throughout the Enterprise
0:45 so people can access their insights and
0:48 uh we were supplier of the year for
0:50 Coca-Cola the year before last uh for
0:52 their knowledge and insights group been
0:54 around for about 20 years
0:59 uh me my name's Jason ruso uh my last
1:00 name actually if you spell it backwards
1:02 is oh sugar
1:06 so uh true story and
1:09 um and uh you know for those of you that
1:11 don't know me I actually have the best
1:15 job in the world I get to work with uh
1:18 clients and uh Executives that have very
1:20 challenging and very complex analytics
1:22 problems and I've been doing it for a
1:25 long time and uh I've had the privilege
1:27 of working with Walmart United
1:30 Healthcare sap helping sap do their
1:33 analytics uh Starbucks uh
1:36 InterContinental Hotels Nestle uh
1:39 General Motors and so over the course of
1:41 the last 10 years and helping to grow mu
1:44 Sigma and helping to uh deliver all this
1:46 uh work I thought I'd compile some
1:50 learnings and uh some of the uh what I
1:52 think are uh unintended consequences of
1:55 people going a little too fast when it
1:57 comes to analytics and where people get
1:58 where people get snagged
2:00 snagged
2:03 actually go back one second here uh so
2:05 the title of this presentation is Arch
2:07 contacting an intelligent organization
2:09 so what is intelligent or how am I
2:12 defining intelligent um you know intell
2:15 an intelligent organization to me is one
2:19 that uh balances uh performance um Eric
2:22 was talking about uh Precision uh you
2:25 know accuracy speed quickness but then
2:27 also the ability to innovate and um you
2:31 know the
2:33 I Big Data you know the word Big Data
2:35 means a lot of things to a lot of people
2:37 um for me I think it's a metaphor for
2:40 disruption I think data's always been
2:42 big uh but I think if you actually take
2:45 the two words big and data I think they
2:48 represent uh two sides of the coin you
2:52 know the the big part is um is the part
2:54 that allows CEOs to come up with new
2:57 business models Coos to come up with new
3:00 ways of uh driving driving efficiency in
3:03 their business CMOS new ways of reaching
3:06 their customers and uh you know that
3:08 part of Big Data you you know the big
3:10 part I think is is represented by Albert
3:12 Einstein's quote it's you know
3:14 imagination is more important than
3:15 knowledge you know that's where the
3:18 Innovation comes from you the data part
3:21 of it to me is uh you know how you
3:24 gather and store and access and uh
3:27 assimilate that information in a quick
3:29 and uh accurate manner I think is more
3:31 presented by you know the W her quote
3:34 you know fast is fine but accuracy is
3:37 everything uh and the reality is for
3:39 just about everybody in this room to
3:40 some degree if you're involved in
3:42 analytics in your Enterprise you're
3:46 accountable to both performance and
3:50 Innovation and so you know how does that
3:53 feel well unfortunately you know speed
3:56 doesn't equal learning and and accuracy
3:59 doesn't equal experimentation and so the
4:01 ability to deliver significant
4:03 performance a lot of times actually is
4:05 you know compromises your ability to
4:09 innovate and so you know this is uh you
4:10 know I think a picture says a thousand
4:42 and so there's performance on the left
4:45 and there's Innovation on the right and
4:48 uh and you know there and we all have
4:49 that look on our face that Homer has
4:51 right there hanging from the wrecking
4:56 ball as uh you know our um as as a new
4:58 technology comes in because what happens
5:01 happens
5:03 it can deliver both performance and
5:06 Innovation and then suddenly it's the
5:10 you know it's incredibly hyped and then
5:19 and and so what I want to really cover
5:21 today is
5:24 is I want to talk about the mindset
5:26 that's required for successful uh
5:29 analytics I want to talk about uh the
5:32 skill set or at least my perspective on
5:34 skill set I think actually par of and I
5:36 have some very similar uh opinions on
5:38 this and then and then I want to talk a
5:40 little bit about a level set um I want
5:42 to talk a little bit about what this
5:50 cultures and so if you think about
5:53 analytics uh regardless of the actual
5:55 industry you're in you know the business
5:58 that we're in as as analytics Executives
6:00 is the business of connecting data to
6:03 decisions decisions to data that is that
6:07 is our business how we do it so we will
6:09 take the outcomes or we'll take the uh
6:11 you know the exhaust the digital exhaust
6:13 from a business environment create these
6:17 data stores and then we'll do
6:20 analytics on the data and then people
6:22 will make decisions and that influences
6:24 the business environment and and this is
6:28 how it goes right and so and then to
6:30 kind of operationalize it you know in
6:32 the lower left corner you have your data
6:34 infrastructure and that's everything
6:36 from your relational databases all the
6:39 way to your Hadoop clusters and all
6:40 points in between and then you have your
6:42 analytics which is everything from your
6:45 descriptive to your interpretive to your
6:48 predictive to your prescriptive although
6:50 I sometimes debate whether prescriptive
6:52 actually is its own thing or if it's
6:54 just an attribute of descriptive
6:56 interpretive and predictive I mean all
6:59 analytics really should be prescriptive
7:02 and say so what and now what and I guess
7:05 while I'm opining I'll also say that I
7:08 think it's kind of a sad commentary that
7:10 in the times we're living in today we
7:13 call it datadriven decision-making you'd
7:15 think we just call it decision
7:19 making uh so uh you know as we evolve
7:20 though and you know analytics gets more
7:22 mature then there's the process of
7:23 insight Logistics and what I mean by
7:25 Insight Logistics uh I think someone
7:27 touched on this yesterday although
7:28 thankfully they didn't call it this is
7:30 you know getting the right information
7:32 into the right executive's hand at the
7:34 right time in the right way so they can
7:37 make the right decision and and it's the
7:40 skills required to properly deliver
7:42 analytics in my opinion are inherently
7:45 cross functional the idea that your data
7:48 uh Architects can be in a silo and you
7:50 know not really you know be in the same
7:52 Department as your your your
7:55 statisticians who aren't you know in the
7:58 same Department as your uh as your folks
8:00 that are actually uh you know working
8:02 with the business to help them uh Drive
8:04 business outcomes I think that that's
8:08 going to have to go away because uh the
8:11 the the speed at which we're
8:15 moving almost incapacitates true
8:17 specialization the way it was once upon
8:19 a time that the ability to be cross
8:20 functional the ability to work across
8:23 all of these uh different concepts and
8:25 for someone in the data environment to
8:26 still be able to connect the dots with
8:28 okay if I store it this way and I want
8:30 to access it this way what does that
8:32 mean for the end user there has to be
8:34 that level of accountability to what
8:37 they're doing but what happens got all
8:39 this great technology and I think that
8:41 uh and part of you know not a silver
8:44 bullet I could not agree more I what I
8:46 do think is important is that people
8:48 need to be able to play around with
8:50 these things and figure out the context
8:52 and how they relate to one another and
8:54 then figure out if it's meaningful but
8:57 uh you know Cass whether it's Cassandra
9:00 or whether it's uh you know spark or
9:03 what have you I think that uh people
9:05 often times are like oh gosh we got to
9:07 have this skill set inhouse before
9:08 they've even really validated whether or
9:10 not it's truly meaningful for them and truly
9:11 truly
9:14 complimentary and so let's go into
9:17 mindset for one more
9:19 second you've got this business
9:21 environment you know and you you've got
9:22 this data and you got this analytics you get
9:23 get
9:27 decisions to me this is actually a
9:29 performance mindset when I think about
9:31 big data and what big data is really all
9:34 about you know especially
9:36 combined I think that there's more to it
9:37 I I think that if you're going to drive
9:39 Innovation you have to think a little
9:41 bit differently you have to say well
9:43 wait a minute you know what decisions do
9:45 I want to be making and what analytics
9:48 do I need to be doing and what data do I
9:50 need to
9:53 have and then can I reinvent the
9:55 business environment and it's our
9:58 Charter as analytics Executives and the
10:00 ones that I work with to make sure that
10:03 they're capable of doing both for their
10:05 Enterprise so how do they do that or
10:07 what's the you know what's the what's
10:09 the way that they do it well I believe
10:12 that it's it's it's Unique to each
10:14 culture and there are um there are
10:17 different cultures there are first of
10:24 adopt well you know some will build and
10:26 some will buy and uh I'm not here to
10:29 pass judgment I think that there's uh I
10:32 think there's a case that you know both
10:34 are are good uh but they also have you
10:36 know every strength is a weakness uh so
10:38 if you build something you obviously you
10:39 know you own it and it's but it can be
10:42 slow uh but over time you can achieve
10:45 certain uh scale benefits you buy it you
10:46 can have it very quickly but sometimes
10:47 you know there's a little bit of
10:54 fast and then you have deployment and so
10:55 you know should I be centralized you
10:57 know having all these data scientists in
10:58 one room there's a lot of benefit for
11:00 that that for all the sharing or should
11:02 I be decentralized I could be closer to
11:04 my customer I could be more responsive
11:07 and you know what happens is companies
11:11 tend to wind
11:14 up falling into buckets and so
11:17 centralized companies that uh build
11:19 their own Technologies or their own
11:21 capabilities you know they have great
11:23 structure they're
11:25 slow and uh you know for folks that want
11:27 to be
11:29 Innovative I I hate to break to you but
11:30 if you're come from this quadrant you
11:32 are going to be challenged in driving
11:35 Innovation you will be out of this world
11:38 when it comes to driving performance but
11:39 you know you every strength is a
11:42 weakness right and then you have the
11:44 outsourcers so you know centralized you
11:45 know they tend to align around one or
11:46 two technology companies you know
11:48 they've got they've got IBM they got
11:50 Oracle and then you know fill in the
11:53 blank for the third one and you know
11:56 maybe uh TCS and and they wind up
11:58 Outsourcing a lot and they go very fast
12:01 in some regards but there's also
12:03 sometimes uh you know it's it's a lot of
12:06 expense of course and they're current
12:08 but um you know they can be challenging
12:10 sometimes because uh they've outsourced
12:12 a lot of capability there and so when
12:14 new technologies and New Concepts come
12:16 in they have a hard time uh really
12:18 knowing what that means they have to
12:19 just listen kind of to what their
12:20 Partners tell them and there's some risk in
12:21 in
12:24 that and then you have the Islanders and
12:26 I see this one probably the most in the
12:28 companies that I work with frankly well
12:29 probably the bottom to
12:32 but um you know the Islanders are um you
12:33 know so they're
12:35 decentralized uh and uh you know they
12:36 tend to build a lot of their own
12:38 capabilities I see this a lot in cpg
12:40 because multinational companies they
12:42 want to be close to their markets uh and
12:43 so what happens in these kind of
12:46 cultures is uh very very effective and
12:48 uh you know kind of very good at get
12:50 capturing what they need to capture
12:52 within their markets not so good at
12:54 communicating with each other and they
12:56 don't collaborate as effectively as they
12:58 could and so um you know you tend to
12:59 have have
13:02 um communication gaps and uh
13:05 opportunities lost uh in the way that
13:07 they could be sharing better and then
13:09 you have the Trailblazers you know
13:10 they're decentralized they buy things
13:13 they move very quickly and uh and so you
13:16 have um you know very fast but also
13:18 chaotic maybe even I'd call it
13:21 schizophrenic but uh they're actually
13:24 very um you know they're they're very
13:26 passionate about customer service and uh
13:27 and these are usually more uh your
13:29 startups like uh I did a lot of work
13:33 with uh Airbnb they were like this and
13:35 uh and so their challenge is how do I
13:37 how do I deliver you know um
13:39 consistently so I'm not just Reinventing
13:47 time and so the reality is as you guys
13:48 you know leave here today and you take
13:49 all this knowledge and you go back and
13:51 you lead your organizations you're going
13:53 to have to build these capabilities
13:54 inside the culture that you that you
13:57 know is going to be hosting uh these
13:59 analytics so I just wanted to give you
14:01 guys a few case examples and takeaways
14:05 from uh experience and also um uh some
14:11 potentially so how do I drive service
14:13 and and to me like the the controllers
14:14 like I said I find that they have some
14:17 Innovation challenges uh there's a book
14:20 called The innovators dilemma uh it's a
14:21 good one I think it's related to these
14:23 cultures in particular they're very good
14:25 at doing what they do and then then
14:26 that's kind of the optic through which
14:29 they look at everything um I think that
14:30 the the best advice that I could give
14:32 for companies that come from this group
14:34 is to rethink your sandboxes a lot of
14:36 times the you know they'll they'll
14:39 create a Dev environment and then uh uh
14:41 and then they'll uh you know they'll
14:43 invite some power users and uh they're
14:45 like you know okay well we can play
14:47 around check and uh you know we got some
14:49 people from the business the power users
14:51 but more often than not the power users
14:53 in these kind of cultures aren't
14:55 actually that close to the business uh
14:56 you know because they're such power
14:58 users and so I think you got to try to
15:00 invite people into the sandbox that are
15:03 also like almost naive about uh what the
15:06 technology is or what the tools are and
15:08 just hear what they're trying to
15:11 do uh a good book for uh to read is uh
15:13 where good ideas come from by Steve
15:15 Johnson uh it's a very powerful book
15:20 about how um how you know one of the
15:22 examples is you know Newton was thinking
15:24 about gravity for a hell of a long time
15:25 before the Apple hit him in the head it
15:27 was just that moment that suddenly it
15:29 crystallized and so there's a lot of
15:32 good ideas about how you can work within
15:37 your Enterprise to um to learn and uh
15:44 way the
15:46 outsourcers you know how do I instill
15:50 ownership you know
15:52 um so the you know the companies in this
15:55 group tend to throw money at problems
15:58 but they won't throw executive time and
15:59 so there's there's a lot of leadership
16:03 issues there uh as far as commitment and
16:05 so uh my advice to the the the folks in
16:09 this group is um try make sure that you
16:11 build an internal competency as well
16:15 like uh you have to having worked with
16:17 these companies I can tell you that um
16:20 there's a little bit of a a confusion
16:23 about well what am I paying you for why
16:25 do I need to have you know analytics
16:27 expertise inhouse if I'm bringing you on
16:29 and and my answer is that if you don't
16:31 have some analytics expertise and some
16:34 fluency inhouse like like solid then
16:37 your leadership will never uh really be
16:39 able to understand what we're doing for
16:42 you because you know we're delivering it
16:45 through you as my sponsor and uh if
16:47 you're not you know capable of doing a
16:48 lot of this stuff yourself or or or
16:50 being out and Walking The Halls and
16:51 socializing it the right way and people
16:53 understand what you're talking about it
16:55 won't be sticky the way you want it to
16:58 be uh a great book is a vested out
17:00 Outsourcing very good book on how to get
17:05 Partnerships and then there's the
17:08 Islanders so uh like I said lot lot I
17:09 see this a lot in multinational
17:14 companies cpgs uh how do I promote
17:16 collaboration and uh and I'm not just
17:17 talking about even just like in
17:19 marketing analytics across all the
17:21 different continents for example I even
17:23 mean like
17:25 um and uh Parts have use the word
17:28 Federated so I actually like that word
17:29 quite a bit I like to call it the Jedi
17:32 Council but the idea is how do I get
17:34 people that are responsible for
17:36 analytics period uh you whether they're
17:38 in the supply chain or whether they're
17:40 in finance or whether they're in uh you
17:42 know marketing or whether they're in
17:44 digital can I get them all together so
17:45 they can start sharing because at the
17:48 end of the day regression techniques can
17:49 be used across the Enterprise in
17:51 different capacities and so hearing the
17:53 different use cases will spur ideas it
17:56 will also facilitate a lot of sharing
17:58 and um and you can start to do some
18:01 really great stuff uh I've seen it time
18:03 and time again and the the only other
18:05 trick to this if you're going to make it
18:07 work is um you have to have somebody
18:10 that isn't the CIO although it probably
18:12 could be but uh somebody at a very
18:14 senior level that oversees this Jedi
18:15 Council so when they get together and
18:17 they knowledge share when there is a
18:19 really good idea that individual then
18:22 can go to the the sea level talk about
18:24 what we're doing and analytics starts to
18:27 become a true function uh again
18:28 borrowing from part to but I think that
18:30 that's that that's when that starts to
18:32 happen great book to read is group
18:35 Alchemy it's all about how uh you can
18:38 take folks that have separate
18:41 um agendas if you will and uh get them
18:43 working together so they're more uh
18:45 distinctively collaborating and uh
18:47 winning for the
18:52 group and then the Trailblazers uh I so
18:55 personally I probably the way I think in
18:58 my life I'm probably a little like this
19:00 uh and so you know the challenge here is
19:02 how do I ensure scalability and uh you
19:03 know how do you channel that
19:05 schizophrenia a little bit because
19:06 because I because I do think that
19:08 there's a distinct advantage to moving
19:09 fast and these people do tend to move
19:12 very fast but being an employee in this
19:13 environment can be very challenging it
19:15 can feel very schizophrenic because you
19:16 know the priorities of the day can
19:18 change you know from morning to after
19:20 lunch and uh and that could be very
19:24 challenging and um confusing frankly and
19:26 uh so a good book to read is a one
19:29 called traction
19:31 the uh the bias of the author is a
19:34 little bit more to uh eradicate
19:35 schizophrenia I don't know that you have
19:37 to totally eradicated but I do think it
19:41 needs to be channeled uh and uh because
19:44 these you know kind of take it a step
19:47 back you know these groups right here
19:49 the uh where's the pointer right there
19:51 you know this one right here real good
19:53 at uh at performance not so good at
19:56 Innovation guess what these guys you
19:59 know not not good at per performance
20:02 graded Innovation so you know it's uh
20:04 I'm not passing judgment by the way I
20:06 want to make very clear that all of
20:08 these are uh have distinctive strengths
20:09 and weaknesses so I just want to make
20:12 sure that I'm not like it doesn't appear
20:13 that I'm being dismissive of any of these
20:14 these
20:17 cultures so I'll leave you with a quote
20:19 from uh you know a data scientist far
20:21 before his time you know we cannot solve
20:22 problems by using the same kind of
20:26 thinking we used when we created them so
20:29 as you go back to your uh your day jobs
20:31 and uh you think about how I can drive
20:33 the effectiveness of my analytics
20:35 organization you know be mindful that uh
20:38 you're accountable to both the uh the
20:41 performance and The Innovation and uh
20:59 questions oh there we go so
21:02 so
21:13 situ so uh let me just restate the
21:16 question so uh you're talking about in a
21:18 very large company where you'll have a
21:20 group that behaves like controllers and
21:21 a different group that's behaving like
21:25 Islanders um so I think in both of those
21:28 cases uh a lot of times the Oru
21:31 facilitates that and uh and and what
21:34 you're trying to do is often times
21:35 especially when it comes to Big Data is
21:38 to drive more Innovation and so uh what
21:41 I found I'm working with a mortgage
21:43 company and uh what we've done is we've
21:45 actually taken and broken off some of
21:48 the high performing more fixed parts of
21:50 their business and there's a group of
21:51 data scientists that are in that group
21:53 and all they're doing is taking the same
21:56 8 to 12 models that they've built that
21:57 run their business and they're
21:58 constantly trying to drive drive
22:00 performance and accuracy on that and
22:01 then there's a different group that's uh
22:05 not not quite a statistici if you will a
22:08 little bit more uh business user
22:11 oriented and uh and they're accountable
22:13 they roll up to the same person but they
22:14 have different uh metrics by which
22:17 they're judged and by creating that
22:21 differentiation I found that um you can
22:23 kind of Tear Down the Walls a little bit
22:25 H because the the business users they're
22:27 always talking about like what's
22:29 possible and that gets people excited
22:31 about that and then uh once they've
22:32 galvanized around what they want to be
22:34 doing they can then try to
22:36 operationalize it that's that's what I
22:38 found the best the biggest challenge I
22:40 find for analytics Executives frankly is
22:41 that they don't have enough time Walking
22:43 The Halls they spend more time with
22:45 their their um you know their sleeves
22:47 rolled up hacking out SAS code or trying
22:48 to figure out why two reports aren't
22:51 tying out as opposed to like literally
22:53 talking to business users about you know
22:54 what's working and what's not working
22:56 and you know I delivered this model for
22:58 you how come you're not using it it and
23:01 well I don't really believe it okay well
23:02 the math is right well I don't trust
23:04 your assumption so you haven't really PE
23:17 that yeah well or some senior leader
23:36 um I think it so to the degree companies
23:38 have I think there's a lot of people
23:40 that kind of are filling the role of a
23:44 CDO or a CAO these days uh as opposed to
23:46 people that are actually have the title
23:47 of Chief analytics officer and then sit
23:50 at the same level as the CMO and the You
23:52 Know Chief people officer what have you
23:55 um regardless I think the the the point
23:57 that I was trying to make is that it
24:00 needs to be someone that has the ears of
24:03 all of those uh you know operational
24:06 heads you know all of those uh sea Suite
24:09 leaders so when a good idea happens it
24:11 doesn't happen in a vacuum it you know
24:13 it's able to be communicated across the
24:15 organization and socialize so people can
24:17 get more excited about the possibility
24:26 enables do that make sense
