Coffee Talks

New Feature: Roundtable

Welcome to an experimental new feature here at P3 Adaptive , where members of the team discuss various topics related to Power BI, Power Pivot, and Analytics/BI in general.  These conversations take place during the week on our Slack channel, and are then lightly edited for publishing on Friday.

This Week’s Topic: Better IT/Biz Relations, Data Access, and “Selling” Power BI / Power Pivot Internally

ROB: OK Austin and Kellan, we’ve been dying to do this for awhile. In your all-too-recent previous lives, you found yourselves championing Power Pivot and Power BI within your (former) companies, and found that the orgs around you were slow/resistant to changing gears. This “insider’s view” of cultural pressures and obstacles is a relevant topic these days for MANY readers, so kick us off.

AUSTIN: I formerly sat in the business user seat as an Excel pro, where I discovered the brave new world of Power Pivot and Power BI. Despite the obvious (to me) power of the next-gen tools, it was difficult to get others on the same page. Both IT and Biz management were reluctant to change gears. I’ve subsequently wound up here on the P3 Adaptive team where I don’t face those politics directly, but our clients do, and I’m sure there are many others out there facing the exact same challenge.

KELLAN: Yeah, Austin and I shared a common experience from our former cubicle-bound existence. Tell me if this sounds familiar … We both started in the business as analysts, as “just Excel people,” then we discovered Power Pivot and gained super powers. Pretty soon we weren’t just Excel people anymore – other analysts looked at our work with reverence and awe. And with our new abilities we had taken a big step toward the IT people, but we definitely weren’t IT people. We were in this uncomfortable “twilight zone” halfway between biz and IT.

AUSTIN: We had wandered off the organizational map – we weren’t this and we weren’t that. It made us powerful, and we could see both sides (IT/Biz) clearly for the first time, but at the same time we were now outcasts, and our organizations couldn’t figure out what to do with us.

ROB: I can certainly believe that, even though I’ve never personally lived there myself. It’s particularly easy to see how IT wasn’t stoked about this “rising force” in the Biz, but I’m getting the sense that you also felt unwelcome on the Biz side? That’s harder for me to understand – shouldn’t Biz mgmt have been coveting you rather than rejecting you?

AUSTIN: Well they (Biz mgmt) weren’t rejecting us, we sort of rejected our limited roles.

KELLAN: It was weird actually. Biz management didn’t have any understanding of the value of what I was doing, and started to think I was just… “technical.” When really I was just a new breed of analyst, but they couldn’t see it clearly yet.

AUSTIN: …and Biz managers were back channeling with IT, and IT was saying that we didn’t know what we’re talking about.

ROB: Ugh. So… Biz not appreciating the magnitude of the new capabilities (and not appreciating YOU properly, I’m sure), and IT actively resisting.

AUSTIN: Continuing on that front, it was hard to understand why the IT team was so reluctant to give me the data my boss was asking for. I felt very stuck in the middle.

ROB: Here are the traditional reasons why IT keeps data “locked away,” in my experience – let’s see if this matches your view from the inside:

  1. Performance concerns – direct query against operational systems can bring things to their knees
  2. Accidental misuse – “you don’t understand how the data is stored, so you are likely to make mistakes and I don’t have time to walk you through it.”
  3. Security concerns – the system isn’t set up with granular permissions, so you will see things you shouldn’t.
  4. Just general fear and desire to control things.

I think we can all agree that there definitely IS something legitimate to 1-3. And yet, psychologically, the Biz usually ends up thinking #4 is the biggest reason. I think that happens because the Biz (rightly) perceives that IT doesn’t factor the *benefits* of data access into the equation, only the downsides.

AUSTIN: Yeah, as a business user it’s hard to fully appreciate how the IT department’s thinking about this.  I remember feeling that people were saying no just because they could (#4) – I wish someone from IT had been willing to explain their full reasoning. I would have understood reasons 1-3, but they probably aren’t used to the business “getting” stuff like that.

And hey, shouldn’t a Power BI “new breed” Power Analyst get taken a bit more seriously than a traditional Excel analyst?

KELLAN: Don’t we like data warehouses? All the data in one place, data’s getting refreshed and cleaned on a regular schedule, the business can just kick back and consume the data. When I was in the business, I always went to the data warehouse first, then I filled gaps from other data sources as needed. I certainly acknowledged the value and cleanliness of the EDW and wished I could have used it more. The problem, generally speaking, was any incremental change required an unjustifiable amount of time to implement. It’s my data and I want it now!

ROB: Yeah we love a good data warehouse BUT they’ve historically been so expensive, time consuming to build, and to your point Kellan, slow to update. Plumbing is just expensive and therefore very hard to justify to management – “hey we want to go spend 5-7 figures on a plumbing project, but take our word for it – THERE WILL BE FAUCETS SOMEDAY.” Good luck with that pitch.

And now’s the time to invert the equation – biz needs a faucet, the faucet has real business value, and the faucet can be built with a minimum of plumbing. Everyone wins, the business gets the information they need and the IT teams as long as they’re willing to accept that an incrementally-built data mart is a win (as opposed to the full-blown warehouse of their dreams).

ROB: Changing gears, let me suggest a few things that we have going in our favor with the new tools:

  1. The new tools (Power BI, Power Pivot) do NOT require all data to be in a single place, unlike previous BI tools.
  2. They allow us to build data marts in “reactive fashion – since we can iterate in near-real-time in Power Pivot and Power BI, that quickly tells us what we’re missing in the DW/DM. And usually, the thing that’s “missing” is really just a quick change – that can be added as a view in a few minutes.
  3. So rather than start from scratch and try to plan and build a “be all, end all” data warehouse, we can create a copy of the operational data source, start using it as a data source, and iterating from there. It’s shocking how much faster and cheaper that is, and yet… it works.
  4. Biz and IT want the same thing, and for the first time it’s within reach, but they need to learn to speak a common language.

 

It’s this common language that’s been missing and it’s not really anyone’s fault – because the older tools (traditional BI, traditional Excel) have been holding us all hostage.

AUSTIN: It’s hard to see it all the time but if both sides were willing to adjust their philosophies and tool sets *just a touch*, they could meet in the middle and get all those human things we want. Not to mention, major WINS for the biz – ROI, profits, savings, you name it.

ROB: One specific thing that we both (IT and the Power BI analysts) want, in terms of tech, is to have reliable data sources: trusted, supported, and blessed – at least for the central biz data – that we can access.

Once you give the “power tools” to the biz, suddenly their manual effort drops dramatically, and the time they spend cobbling together data (via end runs around IT in some cases) becomes a crushing percentage of the drudge work left over – whereas before, it was only a sliver. The net result is that biz is, for the first time ever, incented to the table to ask for “good” data sources. This is IT’s chance to control things like they’ve always wanted, BUT… they have to bend just a bit. Don’t aim for perfect DW’s. Just copy the operational set and go from there. Sounds like heresy in some quarters.

Oh, and we also both want “one version of the truth” publishing and sharing of the final models and dashboards.

Sorry guys, I’m hogging the “mic.” This stuff just gets me so amped up.

AUSTIN: It may continue sounding like heresy but we’re in a good place, a good time to begin changing things. Rob, your post this week is about manners and treating people with respect, despite differences of opinion, and that’s a good start. It’s fun to have a good enemy in the IT department and we’ve tapped into that feeling – but we also have to help people see the other side of it.

We know Excel & Power Pivot because on the business side we got backed into a corner, were asked to do more with less, and found a free tool already installed on our computer. The IT teams of the world never have sufficient budget to execute on all the needs thrown at them. Excel people go one step further and have “zero” budgets. So… “do more with less” is a unifying constraint in this case.

KELLAN: Yeah lack of budget is ok. Lack of budget turned me into a super mutant.

AUSTIN: Ha, me too. If we had gotten a well-funded, well-staffed data warehouse we’d still be in our former jobs.

KELLAN: I remember reading a post by Matt Allington, a couple years back while I was right in the middle of this at work. The post talked about this big opportunity in relation to this person and the value they have in every company, regardless of size.

This person is completely unique and has no suitable role in modern org’s. It’s very difficult to write this person’s job description because it ultimately squirts back and forth between the BI Director and the Director of “Business Unit” as to where this job should belong. It doesn’t matter where it belongs. It does matter that it exists and that the organization embraces it openly. Any data driven org should rely on these people way more than they do today.

Speaking of Matt…

MATT: Oh hey did someone mention my name? Let me jump in here. I spent 15 years in the business struggling to understand IT, and 10 years in IT where I learnt to understand them better. These two groups definitely don’t speak the same language. Business doesn’t know what it needs, or how to ask IT for it. IT knows what can be done but doesn’t know how to explain it to the business in a way they will understand. And yes, WE (the new breed, the Power BI hybrids) are successful because we cover precisely that in-between zone, and we can do it all, despite the shortcoming of the two traditional factions.

To sell to IT, you need to explain why they will get less beaten up, have more time to put out different fires, and get less complaints from the business. To sell to the business, you just tell them they can have it quickly and it can morph and change as they learn, and as the business changes.

ROB: Nice, yes this brings me to a short list of The Benefits to IT When You Play Nice With the Power BI Slash Power Pivot Mutation:

  1. You (IT) won’t have to be the bottleneck for every single report request from now on.
  2. You’re much more likely to get budget and support for your DW dreams, with the slight twist that they will be built incrementally/reactively
  3. You will finally have insight into what’s happening out there in Excel land – all of those skunkworks spreadsheets that you don’t see (but are somehow responsible for when they break), will now be visible.
  4. You can absolutely control Biz access to data, because they NEED the clean refreshable source (cuz manual effort now sucks 10x more than it used to, proportionally)
  5. The biz apps will no longer be written in a language that refuses to scale/translate to “real” servers. It’s written in SSAS, you can take it over whenever you want.
  6. You will get to manage data sources and infrastructure, as well as the most-critical enterprise reports and models, while shedding many other relatively noisy/thankless jobs.
  7. You will get to be a part of WINS. Big wins. And the Biz will be implicitly (and explicitly) acknowledging that fact – to you, to each other, and to execs – all the time, because it all runs on your infra.

We’ve had great success having specifically that chat with IT, because if you treat them as the overworked and underappreciated human beings that they are (just like the business analysts!), as opposed to as an obstacle, well… most people respond well to that. The biggest challenge is getting to the point where we can actually HAVE that chat.

Feedback

Did you like this?  Find it valuable?  Have anything you’d like to suggest for a future week?  Or hate this and hope we never do it again?  Let us know in comments ok?  We’re here to please!