The Politics of Data – Ep. 284

When a report is wrong, it’s tempting to blame the model, the refresh, or the tool. But most “data problems” are really organization problems: unclear ownership, competing incentives, and definitions that were never fully agreed upon.

In this episode, Mike, Tommy, and Seth unpack the politics of data—who gets to define metrics, how teams negotiate control vs. speed, and why governance only works when you treat it as a designed system (with clear decision-rights) instead of a document you email once and hope everyone follows.

News & Announcements

• Down and Dirty with the Politics of Data — The article the team uses to ground the conversation around incentives, ownership, and why alignment is the real scaling constraint.
• Submit a topic idea (Explicit Measures Podcast) — Send questions, scenarios, and episode ideas for future episodes.
• PowerBI.tips Podcast — Subscribe to the show and catch past episodes.
• Power BI Theme Generator — Create a consistent look-and-feel for reports without hand-tuning colors and fonts every time.

Main Discussion

“Politics” isn’t just office drama—it’s the practical reality that data is shared, risky, and valuable. The moment analytics becomes important, different teams will (rationally) want control over definitions, access, and priorities.

Key takeaways:

• Define decision-rights up front: who can change a definition, who approves exceptions, and what the escalation path is when teams disagree.
• Make ownership explicit (and multi-layered): platform ownership, data-product ownership, and metric ownership are different—name them.
• Governance has to be usable: if the process adds friction without obvious payoff, people will route around it and you’ll lose consistency anyway.
• Align incentives with accountability: quality improves when the teams closest to the source systems own outcomes—not when BI is blamed for reality.
• Separate shared assets from domain meaning: central teams provide standards, reusable models, and guardrails; domain teams own interpretation and usage.
• Write definitions that survive edge-cases: document intent, caveats, and “what counts / doesn’t count” so the metric stays stable when things get weird.
• Treat alignment as ongoing work: org charts change, priorities shift, and definitions drift—plan for a cadence of review instead of a one-time kickoff.

Looking Forward

The fastest way to reduce rework in analytics is to make ownership and decision-making explicit—because clear alignment beats perfect tooling every time.

Episode Transcript

0:30 n hello good morning welcome back to the explicit measures podcast with Tommy Seth and Seth and Mike good morning good morning it’s just a morning good morning it’s a great morning welcome to the morning to you Thursday yes from a snowy Northern part of the United States so hello and welcome from that we are super deep in snow at this point in time so my back is extremely sore winter has finally arrived in January so we’ll have have

1:00 arrived in January so we’ll have have snow till April May like yeah so we’ve just next January pivoted two months into the season yeah it always feels like the weather stores up the bad weather and like it’s nice nice nice nice it’s it’s comfortable no snow everything’s nice and all of a sudden it’s like okay I’m ready now and then here comes all the snow and all the disaster that comes with I like to think of myself as the even killed guy keep my cool in most most things right and fortunately I had a mishap I

1:32 and fortunately I had a mishap I apparently just found out about seafoam for your snowblower did not test mine out this year and I have a longer driveway because I have a disconnected garage B yada yada yada snowblower didn’t work we had a bad shovel and let me just put it this way my shovel is into pieces now oh no yes Tommy got a Tommy had a that was not a level-headed decision on the not a levelhead but you decision on the not a levelhead but what it was getting a the Italian know what it was getting a the Italian passion coming out under the shovel it was a mafia decision

2:04 was a mafia decision so now I have a shovel with the metal tip which helps a time it does help very much so today’s topic real quick before we get into it we’ll go into news here in just a moment today’s topic is called the politics of data and we have an article for this one today that we’re going to go through together I think this is a great article and very relevant for organizations trying to think through how does politics and who gets to own what in the data space work for the team so that’s our topic

2:34 work for the team so that’s our topic for today any quick updates or notes or any items you want to talk about from a just news standpoint there hasn’t been a whole lot out there on the powerbi blog the there’s the switching there’s the what reads to me like a hey guys we need to put something on the blog what what is it what is somebody doing there’s there’s a feature that’s been out there for like

3:04 feature that’s been out there for like HTML on the dashboard that yes was feature enabled and now it’s preset defaulted to not enabled because it’s security hole so there’s that but cool I think they’re shoveling snow too maybe maybe so I I do I do feel like there is a lot more lot more conversation around just data Factory in general it seem so as your data Factory has been a standalone product

3:35 Factory has been a standalone product the P the data Factory gets a pipeline a pipeline can run and do various activities in a in a sequence and it’s very it’s very robust it’s a good orchestration tool as well as doing some data movements yep I’m seeing a lot of blogs here in January just talking about just general working with the the warehouse and data factory data Factory using to orchestrate notebooks I think this is very healthy cuz this is one of those aha moments that people need to understand so I I hope that Microsoft’s continue investing

4:06 hope that Microsoft’s continue investing in the data Factory pieces of of the world or the pipelines portion of fabric because I think actually this is fairly strong and really needed especially when you’re trying to load a lot of data in from someplace and you don’t really want to do it with a a data flow which I found data flows are okay but when we start talking like large scale datas it it’s a little bit different one other or news item here do you guys follow a gentleman by the name of of MIM on Twitter he was recently hired by

4:36 MIM on Twitter he was recently hired by the Microsoft team know I’m talking about gonna have to give me a little more context so MIM is an individual who’s been very vocal about this database like an open source database called duck DB he does a lot of performance testing around like Fabric and Google and other systems and does a lot of performance testing he’s recently moved on to the Microsoft cat team and it’s been really fun to follow his Twitter thread because he’s been tweeting and I think maybe he that on LinkedIn I’m not sure exactly where it is but he’s done a really good job of

5:08 is but he’s done a really good job of just saying look how fast this is I think I saw a blog post the other day he was copying 28 gigabytes in like under a minute by using a copy data activity inside fabric the stuff he’s the stuff that he’s performance testing is pretty impressive so it’s nice to see someone putting it through the ringer and seeing how it behaves see if I can find his Twitter handle here a copy data activity via like pipeline pipeline yeah interesting I’ll have to go find

5:41 yeah interesting I’ll have to go find I’ll see if I can stumble across this article that has been gig how many files or are we just or are we just copying data from one structured table to another that I could believe I table to another that I could believe maybe maybe it’s something along mean maybe maybe it’s something along those lines but it it was yeah that makes it was fast regardless It’s it’s just I don’t have when mean it’s it’s just I don’t have when you get things like Fabric in your hands you get to see it run you can play with it but unless you have someone testing like okay how long how long does it take to move 50 gigabytes you don’t

6:11 to move 50 gigabytes you don’t know like yeah if we’re talking to a SQL Server how long should that take you Server how long should that take people always ask me as a know people always ask me as a consultant like hey how much is going to cost me how long is this going to take and a lot of times I’m like well I have some data points that I’ve used in the past and but I don’t I don’t have like a library of like okay here’s my statistics for every single job I’ve yeah so it’s nice to have meim kind doing some of those things honestly that’s the great part about the community is I’ve been seeing a ton really since fabric came out of people testing and they’re showing the cost or

6:42 testing and they’re showing the cost or the usage on all these different scenarios I know Bernard’s been really good with that there’s been a few other people some a little more negative than others on on their take but at the same time we’re getting a lot of preview and this has always been Microsoft’s Community yes this is to me why adoptions always been so rapid is because you get people who are sharing man was Twitter out when power ba first came out or am I showing my age here what well I don’t know if I

7:14 here what well I don’t know if I remember like when powerbi first came out and then that growth or seeing all the stuff I had to learn you can learn a lot of stuff on the social media platforms now or at least get a lot ton of of information I would 100% agree with you and I I’ll Echo here so that I put it Mims Twitter handle there on there you f I fixed my mistake I’m not following MIM okay you need to follow you need to make sure you follow MIM and again he’s now hired by

7:46 MIM and again he’s now hired by Microsoft so you’re getting a lot more of like broad tests on top of powerbi things and fabric related things anyways sweet highly recommended good follow there and a lot of good learning that comes from here’s his link profile as well so I’ll just put that out here as well go follow him stay up to tune stay in tune with what he’s doing what he’s copying but anyway some really interesting things coming from fabric seems to be very performant pretty impressive stuff anyways I like that stuff moving on to our main topic today Tommy give us a little

8:17 today Tommy give us a little introduction what’s the what’s the gist of our topic today we’re talking about down and dirty with the politics of data an article written by Ralph this is we’ve talked about so much about data culture and the need for data to become more consumed by the organization and this article does a good job of or to me I think it’s a great talking point of when it goes wrong or when it gets instead of being ripe gets a little sour and I I’ve seen this I’ve been I’ve been a part of

8:48 seen this I’ve been I’ve been a part of this where rather than data being a part of an asset or being an asset seen as part of how we make decisions it’s more for control it’s more for how we do things and it’s either business technology teams owning it and feeling like they have control so therefore they have power or organiza or departments feeling the need of don’t share my data because I will share my data because I know how it’s calculated and you get truly the politics and you get truly the

9:18 truly the politics and you get truly the selfishness around their data because they don’t want to be judged wrongly about this and this goes from the top down but it’s both the ownership of that data whether it’s the the flow of it into the organization or how it’s measured and how it’s communicated that how when it goes really wrong and I think all too frequently how easy this is to happen within within organizations there’s a lot of really good points in this article let’s just

9:48 good points in this article let’s just kick off with like what the main headline of the article which I think is a really good observation make no and this is reading from Ralph’s initial like headliner here make no mistake about about it data data is a critical

10:01 about about it data data is a critical business asset therefore control of data is power where there is power and control politics Will Follow I think that’s very insightful sounds a bit like Lord of the Rings it does do we hold the data such a such a clickbaity title there should have been like and click here for the next page where where it gets reasonable see shock 10 shocking images of data

10:32 see shock 10 shocking images of data shocking top of top of the first paragraph the control of Data Insights within an organization will not necessarily play out as power struggles and political battles with the result being one person or team ending up the winner or the loser but it certainly can and along the way the organization as a whole may end up in the loser column that’s a it’s a great opening statement I think I think the only other point I’d make with the article is it was written in 2020 right so powerbi was out there’s no mention in that but the the other aspects I want to dive into as we’re talking through this are

11:05 into as we’re talking through this are are there point in here that fabric is challenging in in this structure it’s a great a great idea so but I I agree right so far with with the the layout of I think there’s an early Point made in the article talking a lot about just businesses want to own their data and I do have some examples of businesses that I have seen owning their data and and not not I would say not manipulating it but then putting their Department in the best light that they can based on their

11:36 best light that they can based on their reporting I have definitely seen things especially in supply chain pieces where it let me say it this way a lot of companies incentivize based on performance the data of that performance dictates some bonus or justification for spending more or not spending more or something like that around there’s there’s usually a a goal of the organization an Enterprise goal is tied to something that is related to data interestingly enough the teams that

12:06 data interestingly enough the teams that are working on the data are the teams that own their own goals so for example a manufacturing plant has a goal to produce something in a certain amount of time well the numbers get just adjusted so that oh we we forecasted 100, 000 units but we only made 85 so we’ll just adjust the forecast retroactively and so now we make our goal like we now we now have met the forecast or whatever that may be but there’s there’s adjustments being

12:37 be but there’s there’s adjustments being made so that way the business unit has the I guess meets the goal right and I don’t think that really helps a lot of times I I’ve seen so many times where the forecast is being adjusted based on reality I I don’t think that’s right I feel like you should leave the forecast alone and that is what it is and the forecast and whether you make it or not it’s that’s but there was there was a there’s cultures out there that allow this to happen and so the ownership of

13:07 this to happen and so the ownership of the data how the data is being massaged what’s filtered in and Tomy you I think you’ve experienced you’ve voiced this in the podcast before very similar behaviors where a team doesn’t want to have their sales numbers down by for whatever reason and they adjust well we we need to take this data out because that was an account that’s not our account anymore and that’s that other team’s account right we filter that out remove that from our numbers because it one it made them look bad but then two now now they can meet their quote unquote goals for the year well I think you to the magic word

13:37 year well I think you to the magic word it’s culture here that’s really the core issue and to your point I don’t think people are trying to own the data for the sake of power or they’re they’re seeing in that light like hey we need to have the power and the control I think it’s either a a Fear Factor in the sense of hey if we’re not reporting the right numbers or someone sees these numbers and they’re not with they expected like that’s going to look take us in a bad light to your point but it’s almost the other way from a from a negative point of view where rather than

14:07 negative point of view where rather than we’re going to show us doing so well well if we don’t show that we’re doing well someone’s gonna get fired or someone’s get in trouble so we’re gonna have this ownership of we need to own this side I I really see it two core areas of ownership there’s the ownership of the technology of how the data is actually coming in and there’s the ownership of how we’re measuring it that I I’ve seen be the two core issues of control and how the data is massaged I like the I like the the middle section of the article where it

14:37 middle section of the article where it starts talking about like the control of data and it start talking about some of the politics come out of who can control the metrics the kpis because those are often the ones that have influence around those numbers that this is what a measure of success looks like so if you’re if you’re deciding on what that looks like and that that seems to dictate a lot of the control around what the data is saying you make an interesting point

15:09 saying you make an interesting point that does I think Mike that falls into this this first aspect of control of data and one that I I can resonate with and probably why sales and Commercial is an easy butt of our jokes around data but it’s not necessarily the the the goal setting of the organization isn’t necessarily wrong but I have seen where especially in commercial Realms if this is our number

15:39 commercial Realms if this is our number we’re going to go hit that number that that Fosters a culture of go get that number at all costs interesting and when you own the data sets of the revenue generating products or services or whatever you’re selling in your CRM tool right that we all know this very well accounts opportunities we’re closing opportunities we’re following up with customers etc etc what what I have seen is not all of those systems are

16:10 seen is not all of those systems are directly tied to the costs of the organization observation what you’re selling yes and the problem that happens is when you instantiate a culture of go get it at all costs without the checks and balances of ensuring that what you’re selling is is meeting the guidelines of your margins then bad things start happening because folks will sell anything at any price to hit the number and that’s where

16:40 price to hit the number and that’s where I think total control of an area isn’t necessarily representing bad data or representing that they’re not hitting their goal but that’s not the whole picture right and that’s where you start to get into account like Revenue operations where accounting like the things of things of finance that round out that whole picture that should be part of the conversation right out of the gate or should be built into those systems so that as you’re chasing that number the

17:11 that as you’re chasing that number the accoun exec doesn’t need to worry about or the organization doesn’t need to worry about not hitting thresholds that you should with every sale to ensure that what you’re selling ticks all the minimum boxes of of pursuing that number well and it’s not just Finance data either think about all the different departments that don’t directly have something tied to finance to me I’ve seen this come Delight all too often when we have when powerbi becomes or the central port

17:42 becomes or the central port we’re going to do reporting and the problem is when we have assumed data think of marketing where they have all their different metrics that they’ve been communicating to the business on our open rate and here’s our or a call center on our call rate and all of a sudden power business intelligence takes over over powerb like Hey we’re going to report on this oh your numbers 20% open R like no no it’s always been over 60 so how could it possibly be whatever number that you’re reporting it’s like well we’re taking these factors into account and they like well

18:14 factors into account and they like well we don’t count that like why and those types of conversations I’ve seen it more often where the control happens is when there’s been historical faulty numbers because it’s never really been owned by a single person or it’s never really been owned by a real process it’s been own this Excel file and this Excel file there’s never been vetting for those numbers be it Finance numbers or again pure just departmental numbers once powerbi begins to filter the data in and

18:44 powerbi begins to filter the data in and it becomes centralized those numbers become much more called it accurate but it goes through the same filter people do not like that because it usually conflicts greatly with what’s been assumed in the past I like your I like your point there Tommy and I I want to say I’m I’m resonating with your comment around analytics are now living inside those Excel sheets and there there’s like pockets of little analytics that are occurring that are serving some

19:17 that are occurring that are serving some value but it’s not really part of the broader schema of things and so in the article it talks about the control of analytics resources essentially the people the skills of the people and one thing I think I I will Echo here is I have had two pretty strong memories in my mind around situations where particularly around Master data management pieces where different teams needed to own Master data inside a a

19:50 needed to own Master data inside a a single definition of what products we sell right and there was a lack of engage engagement from the different teams you

20:02 engagement from the different teams you engagement from the different teams think Finance right Finance know think Finance right Finance controls to your point Seth the cost of the product sales and marketing Define the name the value The Branding some of those things engineering controls the dimensional size like how heavy it weighs and all the mechanical pieces of the of the item right all the teams had different properties and some of them were easier to calculate engineering it is what it is but there’s other things like what size of this battery what label how are you going to call like all these things these are other aspects of this that

20:32 these are other aspects of this that were harder to negotiate and it required someone else’s time to fully vet and populate this and so everyone started going well that’s not my problem I’m not gonna I don’t have time for that you go do it I don’t have time for this you go do it and so no one was actually working together to to own their portion of the data there wasn’t clear align M to you own this we’re signing off on it when it’s a problem you got to go fix

21:03 you got to go fix it I think you that that point should be in this article oh it’s it’s there are no owners right yes it’s where it’s where different team it’s the cross collaboration problem with people where nobody wants to take ownership of the full solution and the solution data solution is separated into multiple different Source systems that nobody wants to like give up or right like and that that is also another huge political like I own I I suppose

21:35 huge political like I own I I suppose it’s it’s not it’s not necessarily you can’t access my data it’s it’s right here fine go ahead like I’m not I’m not owning I’m not owning my data I’m not owning the responsibility of merging it with your data right like that’s what you’re describing here oh the other part I want to bring up one last Point here around this one so of the controlling of the analytics teams it does a good job of saying if we don’t have clear roles and responsibilities we’re this will impact our hiring decisions it will

22:05 impact our hiring decisions it will impact what projects get approved it will impact the priority of your projects and it’ll impact the timeline to get any of your projects done and it said I really like this part here there are there the warning sides of some stealth politics are going on around data is you start seeing the hiring of redundant data resources and I’m you jumping ahead here maybe I am I’m sorry I’m going too far control of analytics resources and I’m so far I’m so far behind you want to

22:35 I’m so far I’m so far behind you want to talk about control of data all right well let me I’ll stop then Sor I’ll pause I I think this is I think yeah anyways this is a really good point so keep going back to what you’re talking about Control Data sorry so so one of the things I want to point out is a little bit maybe a little bit down in there there’s a paragraph that I want to get you guys’s feedback on sure do you agree with this in the in the past tense way in which he describes it while not usually nefarious these situations are

23:05 usually nefarious these situations are often born from the early informal days of the company analytics activities dot dot dot when everyone was forced to do whatever was necessary to get their insights this is a thousand perc this doesn’t why this strikes me as as like a interesting like part because it’s in the past tense and I would argue this is very real right now everywhere I

23:35 this is very real right now everywhere I I think this is yes and no to some degree I think some I think it really depends on the organization I feel like there are some organizations that have this is a past tense comment I think in other organizations it’s very present tense I think it really depends on it will vary between companies at least how I’ve observed it I don’t think the technology matters here either it’s not if you have powerbi you’re more likely not to have this this goes back to I think it really goes back to data

24:05 think it really goes back to data culture I’ve seen this with powerbi where different departments have again their own analytical resources this doesn’t happen in a vacuum or hey we had everything working last week why can’t anyone communicate this week this is years and time of people looking at their own things in their own silos their own definitions and how they have labels and hire kees build up over time I’ve seen the most frequent one that I always go back to because it was so apparent was leads and sales and

24:37 was so apparent was leads and sales and marketing but this happens across the board if you have departments and teams working Silo not working towards Universal definitions of how we measure things and how we label things this is bound to happen it’s just like weeds growing without you actually acknowledging it especially if you’re not Cor cross collaborating and especially I think to me and I I don’t want to go to the centralized thing but without some centralized resource of business intelligence or how do we

25:07 business intelligence or how do we measure things without some committee or governance this is bound to happen because every team gets their own culture over time without looking within the organization if you’re working siloed you’re bound to get your own habitual ways and your own definitions yeah I would I would argue that that happens anyway though like every business unit has its own culture they they operate in different ways with their own data I think the the the challenge to that is a cohesive strategy that hits them in the face across the

25:39 that hits them in the face across the organization and lets them know that there are better ways to do things and engages and helps them doing that that’s why when we talk about adoption there’s this big caveat like you get one shot at that because of that reason you’re you’re disrupting a normalized process so I don’t necessarily think like it can be bad but it can’t like there’s also I would expect it right which is also one of the reasons when we talked about this last Tuesday like this is why you do the workshops in the business units right each business unit is

26:10 units right each business unit is different yeah and I was gonna bring that point up Seth well we talked about the workshops and running assessments this is the biggest risk when you run those workshops for them to go terribly wrong if you if you have all these lack of definitions in this rigid control you walk into those workshops going okay we’re going to have these really productive meetings and immediately people are on guard because there’s that lack of wanting to give up and that’s a big part to navigate if you’re not aware that’s already occurring yeah do you guys see a problem

26:42 occurring yeah do you guys see a problem with like so the the the two points I’d make here like first being and the question question with so many thirdparty applications that are used in organizations is it necessarily bad that a particular business Own It unit owns their data in that in that system as long as it’s the source of the data for analytical systems is what I would argue I because I from from a from a analytics perspective or an individual

27:15 a analytics perspective or an individual that is Keen on accessing data from Source systems to be used in a wider context I don’t want to own everything like it’s not it’s not our role to own the the system or the the system of record for a particular product or or thing right from a business unit it’s to ensure that the data that we’re pulling into a larger ecosystem is managed and controlled and accurate and if like I I think

27:48 accurate and if like I I think that belongs in those different business units this may this may be a very copout answer but I’ll give you mine I I think I I’m with you I want that business unit to own their portion like so for example there’s a sales team that’s collecting leads they’re collecting their data inside Salesforce right Salesforce doesn’t isn’t maybe dumping all of its data into a corporate space anyways I’m okay with the team owning that and and managing that as long as we have a handshake agreement to what are

28:19 have a handshake agreement to what are you doing in that system and if you’re bringing that data to the organization when we see there’s problems we agree upon what the problem is and you have guaranteed or you have agreed to going back to the organization and saying I’m willing to go fix these things like for example people are creative data fields are going to be weird they’re going to shove data into the wrong data field at some point or the DAT will be F falsely incorrect or customer will be linked to the to the wrong region as the as the bi team or the

28:50 region as the as the bi team or the central team here I don’t want to be the one fixing all their problems because their data isn’t clean no and I would agree with that I as long as I have an agreement with them that says hey we are going to identify what does the meaning of quality or clean data look like together the idea is you own it and when we have problems here’s the threshold when data gets below this level of quality you will stop building new things or doing work you will come back and someone will be responsible for cleaning and owning that then I’m yes

29:20 cleaning and owning that then I’m yes I’m good with it I I agree I agree with that because that’s that’s the handshake that has to happen agree when when Enterprise systems or systems of data like the analytics aspects of these things start to plug into those you are the data Steward and the subject matter expert of these things here is what we’re we will notify you and let what we’re using in your system here are the dependencies that we have just so here’s

29:52 that we have just so here’s like here’s what we’re using and in return here’s access to things that you didn’t have access to exactly right it’s it’s an agreement on both sides right

30:00 it’s an agreement on both sides right but but yeah absolutely and I think that’s where there has to be that higher level commit across the board because you are going to take 5 10% of an individual’s time but the benefit is twofold one is you’re pulling their their data into an ecosystem that other people can use and vice versa they can access other information out of the analytical systems I think it’s hard to really answer that question in my opinion because there’s levels of ownership we just answered it I’m gonna

30:31 ownership we just answered it I’m gonna disagree then so oh yeah so I’m saying there there’s two levels there’s the ownership of the technology of getting access to the data because the data has to flow in somehow so whether that’s a business technology team or whether that’s from the Department or that’s it that data still has to flow in and that has to be owned by someone that we’re getting the consistent data float in some some standard database or where centralized system the other side of ownership is the business logic one

31:01 ownership is the business logic one of the things that’s in the article is this idea of the black magic metrics and I think those are two separate areas of ownership what’s the logic that we’re eventually and how we’re going to measure this on a quarterly what’s a member count but you still need the data to come in both those areas can arise in a situation of conflict of well we need this new data yeah I don’t I don’t disagree with you but the your one point I would put in the handshake ideally The Source system has a way to incorporate

31:32 Source system has a way to incorporate that business logic into views or something I can access sometimes it does sometimes it doesn’t if it doesn’t then the handshake is more as we extract the data and create these artifacts these objects that are representative of your system we need you here to understand like what business logic we have to input here to make sure our table matches your table and and I think to me that’s just a shifting of where you’re doing things but that doesn’t change the

32:02 doing things but that doesn’t change the responsibilities you do strike the other point I was going to make here though around this control of data that absolutely is rampant which is okay yes we love we need this report here’s all the requirements that you’re asking for and here’s the data set it’s an Excel document well we don’t we don’t want this Excel document we need to access the source system the API well you can’t do that right right and that’s where SI like business units I

32:34 that’s where SI like business units I have seen a lot of around the politicking is some of it’s some of it’s very very I think I think I’ve seen it more in the areas where it it does matter it is more sensitive data that can’t just be proliferated across the board but at the same time like that just means when when you’re pulling in security folks or when you’re pulling in the strategies of how we’re going to extract that that really isn’t a a blocker anymore but

33:05 really isn’t a a blocker anymore but still that’s where the politics part of it will be where there there’s resistance to open access to do the things that are required for analytics because in many cases and not all the time but one of these where it’s a real pain point is can you extract data from systems sure but what most thirdparty systems don’t do is give you a historical record MH crms are a great example of this where it’s like we just looked at the quarter quarter opportunity statuses right and they just

33:35 opportunity statuses right and they just completely went from this to this the next quarter how who did that why what happened right and there’s no trending no tracking of that unless you’re capturing that data on a daily basis like analytic systems do right that’s what’s hugely valuable I think in in a lot of this is you start to see the trends or the changes over time and then identify bad behaviors or patterns that you otherwise wouldn’t have recognized

34:05 you otherwise wouldn’t have recognized etc etc but I se that as a pain Point in in a negative light of data control I think that’s if you think about I don’t know if you’ve seen like the analytics maturity curve there’s like a a curve where you start with like hey I’m doing basic historical reporting and then I’m looking at like Trends and analytics and then I’m providing recommendations and then we’re getting to like using Ai and generate to me tell me what I should do things as you think about that curve I think what you’re describing there Seth is very relevant because as I think

34:36 because as I think about that progression a lot of business units and maybe even organizations are just focusing on what does the data look like right now and I and I see what happens is there are team members that understand that yes I understand what happened right now but I need to compare a state and time of the those to your point sales opportunities right that’s a great one both with the sales opportunities on the first of the month versus the sales opportunities at the end of the month how do we have no sales opportunities halfway through and then what happened

35:07 halfway through and then what happened or how did we get all these extra opportunities or where where did they all go right there’s no trending of that you can only see what it is right now and you and you by not asking the right questions of what you want to track in the data you’re probably not storing the right amount of information or the right kinds like I shouldn’t be just looking at what it is right now I shouldn’t trunk the table and just replace I should actually be storing a timestamp and storing every day the table what’s the status of every day because now I can actually do different analytics and and do better

35:37 different analytics and and do better analysis on top of it so I think as company what you’re seeing there in that example is you’re seeing companies think through wow this was really helpful to do reporting as of right now how do I start doing better an analytics and usually that requires new tools and this is where I think fabric fits very well into this right giving you Fabric in addition to powerbi allows you to collect data in a different way and maybe more of that data that will then change what

36:07 data that will then change what questions or analytical things you want to ask later on well and you still need that that that handshake or the access point of view I’m I’m I’m laughing because I’m thinking about s point about the Excel sheet it’s like hey especially for a business intelligence team coming in like can we see your data here’s our Excel sheet that’s great but can we see actually where you got it from why do you need that why do you need our source data and I think that there’s a big area of conflict there where people they’ll give you some data they’ll give you what

36:37 give you some data they’ll give you what they’ve already created but to say can we get access to your Source system or your login information or whatever the credentials are to access again there’s hundreds of thirdparty tools or thirdparty systems we using now that business intelligence and fabric would still needs too it doesn’t just need the final cleaned version and it’s like that’s what we’re going to start reporting on because we don’t know what you’ve done to this nice pretty Excel sheet with conditional formatting we need to see the row that’s to me a huge

37:08 need to see the row that’s to me a huge part where the conflict begins to arise because like we’ll give you what we’ve already cleaned why do you need to see anything else but if business again this I don’t even think this is an ownership of the measures this is let’s get your data in a right format because early in the article people have just design things off the seat of their pants because that’s the best they could do but we have better systems now and people are built in their their ways yeah I want to I want to I think

37:39 ways yeah I want to I want to I think this is a good opportunity to move into more of the control of the analytics Resources with the people’s portion of this because I think this is where I was going a little bit earlier in the conversation a little too fast I think what you described there Tommy speaks very well to how do when there are stealth politics coming in and it’s the hiring of redundant resources for a business unit like more people to chew on Excel files is what I’m thinking right oh we have so much work to do we need another resource or two and then all you do is you bring them in and you just throw you just throw them into Excel and they’re just chewing through

38:09 Excel and they’re just chewing through Excel files that’s probably not a good use like there’s probably a bigger problem we should be solving to make all of that a lot easier so what’s that kind of that a lot easier so what’s that issue the creation of of issue the creation of Shadow or un I I’ve heard Shadow it all the time the time and com from the business I don’t like the term Shadow it to me that just indicates someone is Shadow it is someone or some team

38:39 is Shadow it is someone or some team that has been unable to do their job and they have develop their own process outside of the purview of the IT team so why did that like don’t call it Shadow it figure out what is why did they have to do to do that why was why do you call the why is this Shadow it thing appearing is it because you it are not providing the data you need to the business you like everything I see in powerbi right now especially with Fabrica incorporating here is I see the line between where it

39:10 here is I see the line between where it used to do the data work and where the business through the work is now continuing to blur more and more and I think about Rob col’s book I think he was like analyze something in Excel or something like that he’s got some kind something like that he’s got some really great book now that I read a of really great book now that I read a while ago but in the book he just describes describes like there’s there’s this moment where when you involve more people in a direction in atically whatever that thing is the more people you add the more communication that’s required to

39:40 more communication that’s required to understand the direction to be at everyone on the same page to get everything moving the same direction and what powerbi is doing is it’s adding more capability to business and you’re you’re involving I think more people to have input around what data is doing and I I think it’s the right approach approach but I think businesses are struggling to figure out what’s the

40:01 struggling to figure out what’s the right blend of how this works does that make sense it does I I think from my perspective in this realm I I don’t think this is the article paints this in this like mysterious nefarious way that like oh yeah we’re we’re creating problems for the organization but I almost guarantee you this is a perspective thing it’s not like the business unit is able to do things that are like nefarious and like the like the business or leadership team doesn’t understand that like there’s they

40:32 understand that like there’s they shouldn’t be doing it I I think this is a problem of how typically business people operate which is a give it to me now mentality every like we have a problem throw about like it’s it’s this it’s the same thing engineering or business intelligence teams get hammered with hammered with literally you see a nail you whack it with your Hammer what is your hammer it’s your skill set what is your skill set how the business runs Excel you access right

41:02 business runs Excel you access right like so what you see that nail what is it going to take more bodies right it’s G take me me requesting this thing oh that’s too slow so we got to do what ourselves and I think that’s where we get ourselves into trouble in large part because they don’t understand what is possible or that there are better more efficient ways to do the thing that they need to attack and this this is very true in data data tasks right virtuous waste is

41:32 data data tasks right virtuous waste is a thing that’s why we as business professionals you as consultants can literally with 90% confidence walk into a business and say if you give me x amount of dollars I will give you that back in the one month I’m here right I just point me to a business unit I’ll find find efficiencies right and in many cases like we’re we’re not out to fire people we’re not out to people but like you how many times you walked in and be like this this person all they do is move data around if we automate this process

42:02 data around if we automate this process that person can do something else right like analyze the data or like give you insights and like increase the revenue of this business area etc etc so I I think this is where Mike to your point that you were ending on as well that I agree with is this is why the focus on data culture are these really large initiatives that we talk about on Tuesday are so valuable because we’re trying to raise the bar of General data understanding and capabilities yes acoss people yep in their business units

42:33 people yep in their business units regardless of what you’re they’re doing with the data control or analytics resources understand the art of what’s possible yes and when you’re doing that that should start from the top so that as you’re moving down any of these requests that start coming up through the organization should be halted or at least tailored into a different direction of like guys why are you do okay okay so you’re going to do this but do this team over here is doing that or did that we already have an initiative going on here we have

43:03 have an initiative going on here we have resources we have some capacity go talk to them so instead of us agreeing to your two more resources we’re going to point you in this direction and I I just think it’s a it’s a byproduct of areas of business trying to execute in the way that they know how to as as quickly as they possibly can which is what we’re trying to do we’re trying to increase the value of the business in our different ways and we’re in a unique position I think because we have the foot in it and the foot in business

43:35 foot in it and the foot in business right so these perspectives are our responsibility in the adoption road map that we talk about all the time and the data culture to raise up throughout the organization and they’re going to be handled and executed on in different ways but this is where you would start to see I think the analytics resources start to be congreg ated in specific areas as opposed to they need a person and they need a person and they need a person right like that that you start to co like build some cohesion and

44:07 start to co like build some cohesion and maybe you then have resources that are those powerbi Champions or fabric Champions that you consolidate into what is your extended team of a Coe or the the working off the top of my head the the the name of those people the champion team right that are pushing the larger efforts into the organization but that’s to me just part all part of this Collective communication that that has to start happening or at least strategy that has to happen within

44:38 strategy that has to happen within organizations for this to work and it’s very much a domino effect too to your point set where the longer that goes on the the more June 2023 July 2023 Excel sheets you have or people working on that over time when you begin to shine the light on this you’re shaking the core of the data itself or the people working on it because like well no I spent 20 hours on this and I’ve done all this time the longer people have been working in those systems or that same process and now you’re coming in odds

45:10 process and now you’re coming in odds are the numbers are going to change if it’s taken over by business intelligence there’s going to be some flux there and people’s processes are going to change the longer to your point that people have just been doing things because they needed to the more you’re shaking their the more you’re shaking what they’ve known to be true and I think that’s a big point of how do you address that even if their truth is based on wrong information exactly still their truth it’s their truth exactly and that’s a

45:40 it’s their truth exactly and that’s a huge part of this I’ve seen I’ve seen people relying on Excel sheets that have errors in them and they’re like well that’s that’s what we’ve been doing that’s how we’ve been running things okay but there’s clearly errors in here can you see where but it’s wrong but it’s wrong so but your point though it’s like like well you have to get people to acknowledge that okay yes we’re going to move forward we’re going to do better on our analytics but like it’s it’s a it’s a challenge because you have to get people to agree like okay yes it was wrong okay how are we going to do it

46:11 wrong okay how are we going to do it better moving forward and again I think that’s where the another side of the empathy comes in because you’re basically telling them hey go back to your boss and tell them the last nine months of what you’ve told them is faulty and all the time that’s a reluctance why they want to change no exactly because we’ve we’ve eaten that Crow ourselves a lot regardless of with this is why we stress business owners like owners of the data because we’re the ones that are going to get hammered if data is wrong in the report and if

46:41 if data is wrong in the report and if it’s because of a course our system that’s that’s why that yeah okay yep that’s on me yeah and so I don’t yeah to me that’s not just power that’s also you’re shaking up everything where you may think of someone possibly getting fired or shaking up our boss not being happy with the last nine months I think that’s just a big part I look at always how many Excel monthly files or reports that they have and my sensitivity level when I go into those types of conversations yeah so we got

47:12 types of conversations yeah so we got two sections to to to crank through I think controlling the analytics technology like from a framework perspective you guys would agree like that that belongs in a central team or or at least is somebody with technical understanding of like what are the requirements right sometimes that can be a SQL Server other times the the organizational data is going to be much larger and needs a much more robust system but I like this section in that section around controlling the analytics technology there is this it’s the non-core part right it’s the noncore that’s hard part because I agree with

47:42 that’s hard part because I agree with you there are core parts of your business that yeah that comes through a central team you’re going to put it in a central storage area server Lake housee whatever you want to call it and and it gets served from there which makes sense because you again the business has has decided there’s enough value behind that they’re going to plan and design it correctly it’s all this other stuff and I’ve actually seen this too like alricks he mentions and then data science tools I’ve seen this as being very much a problem because you bring in the data scientist and they just do things

48:13 scientist and they just do things literally I was talking about a data scientist a couple days ago I was like where do you do your work he goes oh I just download the python libraries to my laptop and do it there I’m like cool love it that you’re doing data sciency things but like that’s not the place to do this because I can’t productionize your laptop that justes it just doesn’t work like that’s not you can do exploratory things and figure out what may be good and what may not be good but we shouldn’t be doing those things in a vacuum we should be thinking about doing them in notebooks inside fabric like that’s that’s where that stuff should be

48:43 that’s that’s where that stuff should be living that way you can pick that up and add it to a pipeline and you can make it you can output that same amount of data every single day because you’re going to you’re going to be doing predictions analytics and forecast people are going to rely on on those numbers that’s why it’s good to have it in a common common place yeah what do you guys think about these non-core Technologies I I think about access I think about alrix I think about people buying other analytical tools not powerbi like Tableau or have other versions of of

49:14 Tableau or have other versions of of software that are around that are just maybe through an acquisition they’re just there yeah I I don’t like them because next next conversation and and and the reason being this is purely from it’s just not good business because from a cost perspective you’re going to be spooling up many different many different subscriptions or different services for different tools and this automatically leads into specialized

49:44 automatically leads into specialized roles like so all of the sudden now I have to have analytics people in different teams that know how to use these different tools and then not only do I have a difference in how data is being used for different business units and that’s fine right but how do you

50:01 and that’s fine right but how do you reconcile all that like if if and when you get to a point which you will at some point as an organization going why do we have five visualization tools and the business going well it’s what what we’re using used to and someone coming in going nope we’re gonna use powerbi what do you do with all those resources that don’t know it what I’m saying like that’s the conflict here that that this results is it’s very easy to say well business Union should be able to do whatever they want to and it’s not just

50:31 whatever they want to and it’s not just data and Reporting and analytics there’s other decision points too with different applications do they can you talk to them do can you integrate them with other things like all of those decisions happen all the time too that are blockers to Better Business right because as these strategies start to come into play there’s there’s cost there’s overhead and then when you when you proliferate that into the people who have to maintain those specialized things that that’s where you’re going to run into a problem at some point I don’t

51:02 run into a problem at some point I don’t think it’s just analytic tools and I’ll just ask you guys real quickly on this to move on but what about the all nonstraight analytical tools Google analytics sale through email campaign information operations software systems that provide a ton of data that again may not be straightforward into the core systems should those also like those also I think need to have that ation here as well I would put them under I would put those under the lens of like core and if you’re I think I

51:33 core and if you’re I think I think there’s a certain size of data or a certain amount of data that’s important enough where you’re like yeah this is important enough we should be talking about that and and this is where I think leadership and again that working team that working team inside your analytics should be really evaluating this like you will be hearing of important data sources this team always uses this data source this team uses uses this program this team uses this third party application I think those conversations will bubble up but you’re right Tommy I think you need to have conversation around them I think you need to bring

52:03 around them I think you need to bring them up I think you need to identify if they’re strategic or not and if they are then you have the conversation okay Central team how are we getting this data in it and it may not need to join with other core system data it may just be let’s just make it easier for our user to get it because we’re not going to let you use all these other tools we’re going to let you use powerbi so we have to get it there two tools that I really really have some conflict with organizations that are spinning up Snowflake and organizations that are using altrix because I feel like a lot

52:36 using altrix because I feel like a lot of snowflake and altrix is capabilities that powerbi already does so they want to use snowflake they want to use altrix to do shaping and Engineering of data and yet they want to use powerbi as well and I’m like well you realize a lot of what you’re trying to do already can be done inside cutting right there right I feel like I feel like you’re Double Di or spending more money because you’re building two things so if you’re going to pick these things right let’s just pick strategically what we’re going to do here and and try to yeah have a plan to figure out do we

53:07 to yeah have a plan to figure out do we need do we really need snowflake does it fit our longterm objectives and I’ve heard really clearly from Snowflake they’re not putting any AI stuff inside their ecosystem so when you’re going to have to yet buy another tool to to run that in in addition and everything else so I think companies really need to be thinking about strategically how do we get our data how can we get it stored at volume and what tooling can we put in place that will allow us to do deeper analytical things

53:39 allow us to do deeper analytical things the nextg of AI and machine learning that’s it’s going to get there I think I think Tommy your to your original Point like the I’m fine and I’m like 100% on board with reporting from the core the system of record if it’s part of the tool right of course they have reporting efforts and needs a business unit and and as a central bi team I don’t want to own all that do we want to like offer them capabilities and teach them that they can extend that whether need there are need there always needs with joining to

54:09 need there always needs with joining to other data right or aggregating and seeing data in a different way or to our time our perspective maybe plugging into our analytical Data Systems that have the historical record right all hugely valuable especially if they know how to use those tools I think it’s the extra analytics tools where I say standardize if you can absolutely standardize if you can because what otherwise happens is those business people either hire or upskill themselves

54:39 people either hire or upskill themselves to learn how to use those tools and then they they don’t rightfully so they don’t want to do it again which they would have to do if you’re gonna say now you can’t use that anymore we’re we’re going to cost save get rid of your three products and we’re GNA have one product now right yeah yep it it it it it creates a challenge is it an opportunity to learn more yeah but at the same time like you think about those cycles of how a business is already investing in solving their problems and you’re you’re saying they have to use a new text act like that’s it’s the add-on part that

55:10 like that’s it’s the add-on part that I think it’s a risk it it’s clearly a risk of yes I agree all right with that let’s do final thoughts we’re basically out of time here on the article Tommy what do you think what are your your final thoughts you’re key takeaways here for this article honestly for me this is something that happens easily and I think without a lot of bells and whistles without you knowing and it creeps up on you and a lot of organizations it’s the easily

55:41 organizations it’s the easily introduction of Technology it’s the introduction of new metrics or different departments and I think without what we’ve been talking about our implementation Tuesdays without a working Team without some centralization of buyin on what technology or powerbi or how we devise and look at data this is going to always creep in and just because you solve it now doesn’t mean that it may creep in organically in a few months from now so I think just being aware of how politics and how data

56:12 being aware of how politics and how data arises is a huge component to Contin to be on top of it Seth any final thoughts yeah I I’d encourage everybody to go read the article there were several aspects that is actually we didn’t get to that impact budget Etc overall there there’s always going to be politics and data who controls what there’s going to be budgets there’s going to be Focus that we all want to be a part of in solving the business problems I I guess where this leaves me is keep if you keep the

56:43 this leaves me is keep if you keep the organizational goals in mind like what are the best choices for the business a lot of this just goes by the wayside and as long as you stay laser focused on that people rally around that message right and I think I’ve seen a lot of positive directional things that avoid a lot of what what’s been brought up in this article but it’s always there so be be cognizant that’s a good point I think in low of my observation for this one would be in Li of fabric and how fabric is changing

57:13 of fabric and how fabric is changing this again I’ve said it from the beginning I’ll keep saying it again fabric provides a lot more data engineering tools to the business user previously we were making models and kicking out reports that was what we were doing and now we have a lot more potential for more Shadow it efforts to pop up and and get brought together in the same tool I think my takeaway is just be mindful to be open and candid about

57:43 mindful to be open and candid about conversation and talking and reviewing what your needs are and how you can help I think I think people are more willing to work with you when you remove pain from their process if if you can make things easier provide more data make it easier to get access to that data we really can move the needle forward and you want to foster a healthy relationship of talking about what the needs are and how the data can serve U the business goals and objectives so be open to Comm

58:13 objectives so be open to Comm communicate more about what you’re doing and what your needs are maybe that’s maybe it’s my final Point sounds good good with that we thank you very much for listening to the podcast we we appreciate your listenership we the only thing we ask for the podcast is if you like this podcast that this was some good food for thought or you learned something new here we’ really appreciate if you would share it with somebody else either on social media or someone at work just spread the word let somebody else know that you enjoyed this episode or maybe even write up a couple ideas that you thought about this and how do you feel

58:43 thought about this and how do you feel about your data culture and how this will work for politics in your company or maybe not because you might get fired so don’t do that so maybe just say you like the article wait what wait what huh Tommy where else can you find the podcast you can find us in apple Spotify wherever get your podcast make sure to subscribe and leave a rating it helps us out a ton do you have a question an idea or a topic that you want us to talk about in a future episode head over to powerbi. com

59:26 amazing no not a few excellent thank you all very much and we’ll see you next

59:56 time [Music] [Music] oh

Thank You

Want to catch us live? Join every Tuesday and Thursday at 7:30 AM Central on YouTube and LinkedIn.

Got a question? Head to powerbi.tips/empodcast and submit your topic ideas.

Listen on Spotify, Apple Podcasts, or wherever you get your podcasts.