skip to Main Content

McKinsey Report on How "Ripe" Various Industries are for IaaS (I disagree about some of these)

Is Your Brain More Valuable Than You Know?

***Update: May 2015

This article is now three years old.  I’ve been gone from my last company now for more than two years, and the world has changed a LOT since then.  We now have MANY options for publishing Power Pivot insight on the web, and I no longer endorse my former company as a hosting choice – we had, shall we say, religious difference over how to treat customers.  That said, the idea of IaaS is more alive and viable than ever.  More tech options for publishing = easier to get into the game.

A Hot New Industry that Could (Should!) Be Dominated By Excel Pros

I read an interesting article the other day about a new type of Cloud Business Intelligence dubbed “Insight as a Service” (IaaS).

If you’re a BI Pro I recommend you read it carefully. If you’re an Excel Pro I recommend you give it a quick skim just to get the flavor of it.  It’s a good article.  Well-written, thoughtful, and imbued with the author’s experience (and as a venture capitalist, his experience reflects a broad cross-section of industries and firms).

But ironically, I believe that IaaS is MUCH more relevant to the Excel Pro than the BI Pro.  Most Excel pros are already in the IaaS business, but it just isn’t called that.


The Case of the Under-Utilized Brain

Here’s a quick story that I think all of you Excel Pros out there will find very interesting.  It’s the story of one particular Excel Pro, the CEO of my last company, Jeff.

Like many Excel Pros, Jeff’s most valuable expertise is in something OTHER than Excel.  He worked for many years specifically in the Consumer Packaged Goods and Retail industries, where his titles tended to be things like “Director of Sales” and “Vice President.”   (If Retail/CPG management were an Olympic sport, Jeff would be expected to medal).  His considerable Excel skills grew in parallel with his domain expertise, as a means to execute his domain expertise.

As his expertise and responsibilities expanded over the years, it was natural that Jeff’s analyses and reports (built in Excel) came to inform not just his own decisions, but those of his entire organization:

IaaS Viewed From the Decisionmakers' Perspective

Does this Diagram Represent IaaS or the Role of an Excel Pro?
(Answer:  Both!)

Domain Expertise Plus Ability to Execute = Everyone Relies on You

I think the diagram above “captures” what an Excel Pro looks like to the rest of their organizations.  Most domain experts, when confronted by the raw data, have no idea how to turn that into actionable insights.  So the Excel Pro in their midst feels like an absolute magician to them.  Even though domain experts know (in some abstract sense) how to get from A to B, the toolset itself (whether Excel or otherwise) is often an intimidating mystery.

“Techies” are the exact opposite – the toolsets are their strength, but they typically lack the domain expertise.  The detailed needs of the business, as well as the million little nuances of how things should and should not be calculated, are critical.

I’ve written before about this standoff – domain expertise on one side, tools expertise on the other – and how traditional BI projects bog down on the “dark matter” of the incredibly inefficient communication between the two camps.  When you pay for a BI project, 99 cents of every dollar is actually spent on your domain experts educating the techies!

But traditional BI projects are actually quite rare.  Few can afford them.  So what normally happens is that your organization comes to rely on a member of the team, a domain expert who has ALSO become an Excel Pro.  This combination of domain expertise and toolset expertise, in ONE brain, is where the world goes for the vast majority of its insights.

Back to the Under-Utilized Brain!

In the earlier phases of Jeff’s career, you can think of one of his roles as being an “Insights Engine” to the rest of his organization:


I think we can safely characterize most Excel Pros’ roles as being exactly that.   While the traditional Excel toolset leaves a lot to be desired in terms of efficiency, security, etc., the combination of domain and toolset expertise – knowing precisely what you need and how to build it – still holds many advantages over BI projects with 6-figure budgets.  This model has been here a long time and it’s not going anywhere.  For good reason.

But Jeff realized something important about his situation:  that the methods he had developed for Retail/CPG would be valuable to many other companies in that same industry.  When he talked to friends and colleagues at other firms, he would often see that they were operating on much more simplistic and inaccurate metrics.  And the data sets were very similar across firms, the domains were the same, and the problems virtually identical.

This is what Wall Street would call an inefficiency in the market.  Jeff saw it as an opportunity.

So he co-founded a consulting firm to provide better insights to Retail/CPG firms.  (It didn’t use the cloud though and was limited to traditional Excel analytics, so it wasn’t quite “real” IaaS, but getting closer).

A few years later, he realized that there was yet another opportunity to “scale up” and become even more efficient.  So he left the consulting company, founded a new company, discovered PowerPivot, met me, and skipping to the end… we started a successful IaaS business for the Retail/CPG industries.

“Magic” for Team –> Insight “Power Plant” for an Industry

Above, I characterized Excel Pros as “Insight Engines” for their organizations.  Well when you start to sell the services of an Insight Engine to more than one company, it becomes more of a Power Plant:


With PowerPivot and a Place to Publish,
Domain-Expert Excel Pros Can Serve an Entire Industry

The metaphor of power distribution seems particularly fitting.  Originally, folks like Edison intended for everyone to have their own backyard electrical power plant, but that was incredibly expensive – it required everyone to invest in duplicate equipment.  It wasn’t until they solved the riddle of efficient transmission over long distances that centralized power plants became possible, and electrical power became affordable on a broad scale.

Wait, I Thought You Worked at a Hosting Company!

I know, it’s a little confusing.  Here’s a quick explanation.  Yes, we have many customers of our Cloud PowerPivot hosting service.  Some of them are using it to provide insight to their own internal organizations, and others are using it to power their own external-facing IaaS offering:


And yes, it’s about evenly split today between people using it for their own internal needs versus using it as a platform to launch their own IaaS offering.

For an example of a Self-Service BI customer running on our platform, [link removed due to 404] see this newly-published case study from Microsoft, covering our customer BMD.  For an IaaS example, see Black Label Strategies.

50/50 IaaS Already?

Does that 50/50 split surprise you?  It has surprised me.  Let me say it again:  fully half of our Cloud PowerPivot customers are using the platform to operate their own Insight as a Service business.

Now, in hindsight, I guess it shouldn’t have surprised me.  After all, Pivotstream originally built out our Cloud PowerPivot platform in order to launch our IaaS biz for Retail/CPG.  It wasn’t until a year later that we started offering the platform to others.

So the diagram really looks like this:


The Company That Shall Not Be Named Offered the Cloud PowerPivot Platform
But Also Runs an IaaS “Side Business” On Top of It

It’s very satisfying, by the way, to be a client of my own platform.  We walk miles in our Cloud PowerPivot customers’ shoes before they ever put them on.  I’m a techie which makes me far too honest to have been a good salesman in my former career, so it’s a very positive change to now say that I trust my own business to the same platform I offer to others.  Everyone who works in software should have this feeling just once.

Closing the Loop

The key takeaways I’m offering in this post are:

  1. IaaS is already a reality and not merely a future trend.
  2. Excel pros already operate IaaS “businesses” out of their cubes and offices today, it’s just not as efficient as the “real” IaaS business is envisioned.
  3. The combination of domain and toolset expertise in one brain is pure magic, and runs circles around traditional BI
  4. The addition of PowerPivot, and a broadcast mechanism like Cloud PowerPivot, dramatically lowers the cost of entry to launching an IaaS business, and does so for precisely the right audience – existing domain experts who know Excel
  5. On net I expect that the IaaS industry will be dominated not by big software or BI consulting firms, but by Excel Pros who have “turned pro” (or “taken their talents to South Beach”)
  6. Our Cloud PowerPivot customer mix at my former company already bore this out, with fully half of those customers operating their own IaaS business
Rob Collie

Rob Collie

One of the original engineering leaders behind Power BI and Power Pivot during his 14-year career at Microsoft, Rob Collie founded a consulting company in 2013 that is 100% devoted to “the new way forward” made possible by Power BI and its related technologies. Since 2013, PowerPivotPro has rapidly grown to become the leading firm in the industry, pioneering an agile, results-first methodology never before seen in the Business Intelligence space. A sought-after public speaker and author of the #1-selling Power BI book, Rob and his team would like to help you revolutionize your business and your career.

This Post Has 11 Comments
  1. FYI, when I click on a link in one of your e-mails, I get an error that the page could not be found at

  2. Great post Rob. I agree that the domain knowledge is the most important thing in BI, but to say that 99% of all cost for the project will go to education is a little over the top 🙂

    1. Welcome back to the blog Kasper! Good to see you slumming around here again 🙂

      OK, maybe 99 is over the top. How about 95? Seriously, I think it’s about that high once you properly account for it. It’s a shockingly large number no matter what, MUCH bigger than I ever would have guessed before seeing how it works one way versus the other.

      More details on this in one of my CIMA posts:

      I’m telling you, not even you guys at MS understand (yet) just how big of a winner you’ve got with PowerPivot. I only know because of what I have seen on “the outside.”

    2. Rob,

      99% is probably too high, but I wouldn’t go any lower than 90%. I don’t have much experience (yet) with BI, but I do have experience implementing ERP. In my opinion, end-user training, business process designing, and spec writing (reports and functionality extensions) take up over 90% of the resources. This might not be the case during the initial implementation, but I believe it is if you look at the entire lifespan of the product. If a techie job function ends up being larger than 10%, it’s probably because not enough time was invested up front to understand the business or too much time was wasted reinventing the wheel rather than leveraging existing functionality.

      1. Thanks Tim. Yeah, I think most tech professionals would heavily underestimate the “time share” consumed by communication in their projects, and not through any intent to deceive. I like the “dark matter” metaphor so much because despite the overwhelming weight of communication, we are self-programmed not to see it. We are so tightly focused on our intent, which is to build something, and not all of the necessary evils: the prep, the fine tuning, and the iterative feedback. Our eyes stay fixed on the goal while we endlessly walk toward it, but in the end we don’t call ourselves “walkers.”

        Funny story: I joined Microsoft at the tail end of the Office 97 development project. When it was complete, one of my managers took the number of bits on the final retail Office 97 cd, which was approximately 5.2 Billion. The entire Office 97 project, therefore, came down to the equivalent of making 5.2 Billion “Yes/No” decisions. With about 500 people on the project and two years in duration, 250 workdays per year, and a 10-12 hour workday, this worked out to:

        Every single one of those 500 people, EACH making a single yes/no decision, once every two seconds.

        This story proves absolutely nothing of course, but it sure is interesting 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *