skip to Main Content

Moving From If to How in Agile BI and Power BI

My 2010 Wake Up Call

Eight years ago, I lucked into an epiphany: IT-driven BI was going to give way to biz-driven BI (aka “self-service” or “Agile” BI). The new wave of BI tools – led by Power Pivot at the time, which ultimately grew into Power BI – gave business subject matter experts something they’d previously lacked: the ability to execute directly.

And while most people initially looked at Power BI’s bold forerunner as a “flash in the pan,” or at best as a way to cut some corners and save some costs, I’d witnessed something very surprising indeed: it WAS in fact less expensive in practice, but also much FASTER, and delivered much BETTER RESULTS than the traditional methodologies.

Faster, cheaper, AND better – the proverbial “free lunch.” And the BI industry, as we knew it in 2010, was gonna go bye bye. It was super clear, and a tremendous gift to be given a window into it that early on. So I’ve been building this company, in one form or another, for the better part of a decade.

I’ve also been “championing the cause” of Agile BI for all those years. And boy were the defenders of the traditional methodology a lot more stubborn in their doubts than I expected! At first it was upsetting to have people bristle at the idea of change, but over time I confess I kinda grew to relish it. It was kinda fun to be edgy. Revolutionary.

I spent so much time in that underdog stance, in fact, that I almost missed it when Agile BI went mainstream as an idea.

Moving from “If” to “How”

If you look around at what the industry’s analysts are saying today (and as always, I highly recommend Forrester’s Boris Evelson), you will see that Agile BI is in fact now taken for granted. The consensus is basically, “of course you have to embrace a biz-driven BI methodology if you want to keep up. Everyone knows that now.”

Obviously it’s super refreshing to live in a world where the experts are validating things I’ve been thinking for a long time. On the downside, I guess I’m gonna have to hang up the “edgy” schtick sooner or later. I’m gonna miss it.

But while the “head” of the industry is now wide awake to Agile BI, it’s clearly taking awhile for that signal to reach the trenches. Recent experience still indicates to me – strongly – that the world is still just beginning to change its ways. We need to switch gears in the discussion.

So today I’m here to say that from now on, when I write about Agile BI – and the implied importance of Power BI – I’m going to be talking about the How, as opposed to arguing with people who wonder about the If.

“OK then, let’s talk about How.”

First, I think it’s important to evaluate where your organization currently is on the adoption curve (and from here on, I’m going to be talking about Power BI specifically, even though most of the concepts would be the same if you were using a lesser tool heh heh).

Stages of Agile BI / Power BI Org-Wide Adoption

Stages of Agile BI / Power BI Org-Wide Adoption

With the caveat that these naturally overlap and blend into each other quite a bit in practice, it’s helpful (in our experience) to identify four different stages of adoption:

1. See the Promise – at least one person in the org has to become convinced that Power BI is a gamechanger. Historically this happened in some hands-on, bottoms-up manner (someone tries it out and is blown away), but recently we’re seeing organizations buying Power BI in a top-down manner too – indicating that at minimum, someone important (often the CIO) has bought in.

2. Win Buyin – one person is, of course, never enough. So now you have to get others on board. This is obviously a challenge in bottoms-up examples of Stage 1, but it’s not exactly trivial in top-down situations either. The Business is suspicious of a message that sounds like “here’s a new data tool, you should use it instead of Excel,” particularly if it’s being advocated by IT.

3. Spread the Success – once you have a transformed a subset of your business into a Cult of the Converted, others are going to want in – and that’s the difference between stages 2 and 3. In stage 2, you were pushing adoption. In stage 3, others are pulling themselves into the game, and that’s what you want! But now you have a new challenge: the challenge of scaling. How do you empower all of your departments, even if they are enthusiastic about it?

4. Govern, Protect, Robustify – in later stages of your organizational change, this pretty quickly becomes your new “problem.” On one hand, biz-created Agile BI solutions are factually less chaotic than the spreadsheets they displaced. But their increased capability – in terms of data capacity, sophistication, and flexibility – does pose some new problems that spreadsheets didn’t. (Ex: at one of our clients we recently saw > 75 active Power BI gateways). I’d still argue, though, that the truly pressing need here is one of opportunity rather than “defense” – the Spreadsheet Jungle simply could not be governed, but Agile BI can. And there are tremendous benefits if you manage to find the “goldilocks” balance between freedom/agility on one hand – and responsibility/governance on the other.

Co-Development: a Term You Should Get to Know

I obviously don’t have space here to lay out everything we’ve learned over the years, but this one is perhaps The Thing. Over the years we found ourselves organically settling into the most-successful patterns of engagement, and “that thing where we sit down with you and just build what you need, together” kept floating to the top of the list. So we gave it a name: Co-Development. And once it had a name, it became easier for us to start codifying the details of it. Polishing and improving it.

In our experience, Co-development is a critical ingredient in stages 1-3 above, and can help make or break your Governance strategy by implication.

Co-Development: A Crucial Concept for Power BI (and Agile BI in General)

An Internal Slide We Use With Our Team

What Stage Are You In?

Do you recognize your org’s “location” within the four stages outlined above? We can connect you with Co-Development support that’s appropriate, whether you’re in stage one and kicking the tires, or in stage three and overwhelmed by demand.  To fit those stages, as well as varying org sizes, we offer Team, Department, and Enterprise Co-Development plans.

After all, this isn’t just something we write about, it’s what we do for a living.

Let’s get started. Reach out to Brandon and Evan at [email protected]

Rob Collie

One of the founding engineers behind Power Pivot during his 14-year career at Microsoft, and creator of the world’s first cloud Power Pivot service, Rob is one of the foremost authorities on self-service business intelligence and next-generation spreadsheet technology.

This Post Has 12 Comments
  1. Governance can sound scary because many folks have worked places where the head of governance was a gatekeeper, and the governance committee could never decide anything so everybody used reports that were in the testing folder. But reading this post makes me realize that *agile* governance is a thing I need to learn more about.

  2. Over many years I built an awesome traditional BI solution (Analysis Services cube) which is still in use and much loved by the team that owns the dataset. For years I’ve been preaching the move to Power BI, steadily gaining traction. That got a big kick along recently when they were presenting their familiar Excel Pivot Tables on Analysis Services shtick – they actually got some (anonymous) hisses of disapproval from the back of the room. They came back ashen-faced and somewhat shaken …

    I think the younger generation who are entering the workforce just expects their solutions to look and work as beautifully as Power BI, be usable from any device, and adapt quickly to suit changing requirements. Its a wonderful time to be active in this space.

  3. I think that your adoption curve’s final point needs to be flexible depending on the organisation. An organisation’s size and the general level of data literacy among its people is hugely important. If the size of the organisation is large but the data literacy is low, I would argue that governance and protection has to come in a lot sooner before you allow too many people to adopt the technologies and concepts at stage 3. You need to protect your organisation from the potential for many, many versions of the truth because business leaders that don’t already have a good track record when it comes to BI will lose faith in the whole concept if they’re presented with an array of conflicting options. I’d make a case that – again, depending on the organisations (but particularly those where the data literacy has always been low) – you need to “robustify” the general data landscape before you let people loose on Power BI, or something like it. We want “agile BI” not “hasty BI”.

  4. Another ‘on-the-money’ article Rob

    The IT old guard who are either in denial of said revolution or just eager to protect their heavily invested fiefdoms will always look to piss on the chips of this. They perceive it as the ungoverned wild-west and merely an extension of the spaghetti Excel problem

    For that reason, how do you react to the following:

    ‘It’s not supported by IT?’
    ‘What happens when you leave?’
    ‘If the user doesn’t like the look of the numbers who do they call if you’re not around?’
    ‘the data is not secure’
    ‘you’re creating off-radar Dev Ops’
    ‘What would auditors make of this?’

    et al.

    I always counter with the devastating arguments of time and cost. “ok, you can either have a solution or you can go the traditional route, but at what cost or timeframe?” I usually present the first unpolished iteration of the solution before I have this conversation, it never fails to command the necessary attention ;o)

    Your article is timely indeed because I’ve finally gotten past the ‘IF’ now need to do the ‘HOW’ properly having been entrusted but look to avoid egg on face!

    I really like your idea of agile-governance. For any guerilla emerging into acceptance, this feels like the perfect card to have up your sleeve, please tell me more

  5. Still having a hard time understanding how you get IT to be OK with a bunch of data gateways sitting on workstations pushing data to the Power BI Service to refresh the reports published there. At some point you have to have their blessing to install the gateway to push data to the cloud and–having been part of enterprise IT–it seems like that creates a data security issue–when you have people publishing reports that contain sensitive information.

    Also, while I agree that analysts will take the time to learn Power BI (and don’t get me wrong, I’m a huge fan), there’s still plenty of people who seem to just want a report delivered to their inboxes–they don’t want to go the Power BI portal or pull up an app to see how their departments are doing (can you tell I’m in government?).

    I can’t help believing there’s still a balance to be struck between traditional, more controlled BI and Agile BI–at least in large organizations with lots of legacy data.

    Someone convince me otherwise. 😉

    1. Ray I think the key distinction we all have to learn to make is this: creation and editing (of reports, models, dashboards, etc.) on one hand versus management/security/delivery on the other.

      Traditionally, both sides fell to IT.

      IT still has to have control over the latter stuff. But the creation/editing – bottlenecking THAT on IT never really worked. (Even when we *think* it’s working, a closer inspection reveals the usage patterns aren’t what everyone hopes they are, and/or there was a MUCH better result being missed because everyone “thought we’d handled it.”)

      I don’t think the world has come to grips yet with the shift to biz-driven authoring (unless you count the absolute reality that Excel is still king no matter how much BI you’ve got), much less developed the oversight mechanisms to make the hybrid work.

      It’s an exciting time for sure. And everything is always on a spectrum. Of course there will always be some things that are 100% IT, end to end. I’m not remotely opposed to that nor bothered by it. What we’ve observed as a company (and I as an individual) over the past 9+ years is that biz-authored provides such DRAMATIC improvement in ROI to the biz, that we can’t ignore it, we have to get ready for it and be in front of it when we can.

        1. Thanks Ray! Glad it was well-received. Nothing is ever 100% binary, but I do tend to communicate in binary terms for emphasis.

          I kinda need a permanent asterisk disclaimer saying “this is all true, but of course there are valid exceptions.” 🙂

Leave a Comment or Question