Remember this guy? Recently, he and I had a heart-to-heart about Excel in the enterprise.
Two weeks ago I posted my reaction to this intriguing LinkedIn ad for IBM’s Planning Analytics:
My reaction was three words: No you can’t!
It turned out to be a popular sentiment, but Joseph Pusztai of CubeWise reached out to me, hoping to persuade me that, indeed, they could make Excel enterprise-ready. We had a great conversation, but first let me share some of my background, and why I reacted so strongly to the original ad.
Enterprises are not very enterprising
What do we mean by enterprise anyway? Most software vendors divide customers into three bands based on number of employees and global revenue, roughly as follows:
Small Business: Fewer than 100 employees, less than 50M USD revenue.
Medium Business: Fewer than 1000 employees, less than 1B USD revenue.
Enterprise software, then, is focussed on these larger businesses, who have complex organizations, and requirements for integrating data and processes across groups. This focus is reflected demands various abilities of enterprise applications, which include: availability, scalability, integrity, security, manageability and – increasingly important – governance.
One other characteristic I would add about almost every enterprise: they are risk averse, cautious and prudent, at least when it comes to IT. Their product lines may be innovative, their marketing bold, their CEO flamboyant, in back-office IT, you’re living the dream in the constant shadow of the auditors: tight budgets, limited scope for innovation and the over-riding need to keep the lights on at all times.
My voyages in the enterprise
I have spent years building enterprise software to meet these needs, all the while sticking to my radical aspirations by chipping away at the edges of corporate conservatism.
At Microsoft, we built SQL Server Integration Services to replace DTS, bringing highly scalable ETL to our platform. We built Data Mining, bringing predictive analytics into the mainstream of database applications. We lived through the SQL Slammer virus. Those were outstanding projects to work on, and we learned a ton about enterprise demands.
One of my last projects was to pick up the PowerPivot client, or as we called it first: Project Gemini. This project was incubated with insight and tremendous drive by Thierry D’Hers, now doing great things at Tableau, who brought his own unique experience from Hyperion and business applications to the case for an Excel-enabled high-scale analytic experience.
After Microsoft, I worked at Qlik, helping them move from a highly successful desktop application to be a true enterprise leader.
In short, I have been ploughing this field for years, bearing the scars (and rewards) that go with it. My resistance to Excel as a enterprise tool is not merely reactionary.
The alternative factor
Looking at planning in particular, I did learn a lot about enterprise planning customers.
Firstly, there are users who want to do their planning within the ERP system. After all, ERP stands for Enterprise Resource Planning, so that would seem to make sense. But ERP planning modules are often cumbersome and limited to the data held in the system – which is often not the case when companies have grown by merger and acquisition and where different divisions often live with different architectures.
Then there are users who are specialized and skilled in enterprise planning. They often want their own bespoke tool. This is the target market of the IBM ad. These users have demanding and specific needs. Allocations (sometimes called spreads) for example: they want to apply an overall budget for marketing and have it divided and allocated to each individual brand in their business by some sort of rule: perhaps related to planned gross income from the brand. They need tools that support this easily and flexibly.
Finally, there are those who do all planning and budgeting in their standard tool. That is perhaps the majority. If that happens to be Qlik, they will want to use Qlik. I had to fight off customers who wanted to turn Qlik in to a read/write planning planning. Most often, however, that tool is Excel – with all it’s flexibility, power and danger.
By now, you can see where this is going. IBM no doubt felt they could reach out to the Excel users and bring them into the market for specialized tools. How have they done?
Digging into IBM Planning Analytics
Joseph, of CubeWise, is one of those guys who has nearly crossed my path many times, but somehow we have never met. It turns out he’s not the guy in the ad! Who knew?! But I had a very enjoyable call as he walked me through the IBM Planning Analytics features.
As it happens, I have known the core technology for years – and likely you have too – under the name TM1 (Table Manager!) first with Applix, then Cognos, then IBM. Now, they call the Excel-integrated toolset Planning Analytics: not the snappiest of brands, but Anaplan was already taken by their big rival. For short, they call it PAx.
The aim of PAx is straightforward: to be the power user companion to the standalone Planning Analytics Workspace, PAw. PAx consolidates some earlier tools into an Excel add-in. In Joseph’s hands, it’s a remarkably agile tool and fulfills it’s goals of mimicking the planning and reporting spreadsheets users may be used to while leveraging the TM1 proven backend for some of those enterprise abilities.
For example, just as when integrating with an OLAP server, in PAx you have true dimensional hierarchies, not just numerous SUM statements. This is essential in cases where you have ragged or unbalanced hierarchies. Similarly, the engine’s Cube Rules are more expressive of real business logic than Excel’s cell-by-cell expressions.
All of this power is available through Excel. But this is perhaps the most important caveat I have. What we’re seeing is a planning system hosted within the Excel environment. For all it’s advantages, PAx has not tamed the risk of that environment, except in so far as users diligently stick to the PAx features.
Staking their claim
The original claim, to which I objected was: I can elevate Excel into an enterprise-class planning solution.
Can you, IBM? Not quite. You can certainly use Excel as the front end to an IT-managed, enterprise planning server. If you stick with the PAx add-in and it’s integrated capabilities, you have an extremely powerful tool, albeit one that is matched by the PAw workbench, and yet less governed too. Rival offerings, such as Anaplan’s Apps platform, also offer specialized power and ease of use. My feeling here is there is still a danger of the Excel features themselves going astray, whether as scratchpad sheets in the same workbook, or work copied and pasted into new documents.
So what to make of the claim about elevating Excel to enterprise class?
No one’s pants are on fire, but you could burn your fingers if you take the ad too literally.
Many thanks to Joseph of CubeWise. He’s an excellent representative of his company, a wise and witty presenter, and a good sport too.
If your enterprise is struggling to establish a culture of analytics, check out the TreeHive Analytics Gameplan, a uniquely focussed, practical and accessible approach to analytics strategy.