Business Intelligence: Cost or Profit?

If I were selling $100 bills for $80, wouldn't you buy as many as you could possibly get your hands on? Or would you say, sorry, my budget for discounted $100 bills this month is only $800?

That is the question posed by Allan Dib in his seminal 2016 book, The 1-Page Marketing Plan. The book itself is outstanding in many respects but Dib's discussion of the fallacy of setting budgets really got the wheels in my head turning. Yes, he was talking specifically about marketing but what if, I wondered, the same concepts could be extended to other areas of business? It's been about a year and a half since I last wrote an article on this blog; these past couple of years have been a whirlwind for Infinity Insight, with our boutique consultancy tripling in size. It's been an amazing ride. But while I have a list of backlogged blog topics as long as my arm, I sadly have not found much time to devote to writing. Frankly, I am not even sure if I have any readers left after such a long hiatus or if I am speaking to an empty room right now. I guess we are about to find out 🙂
In any event, I found the above question sufficiently titillating that I felt strongly I had to explore it further. Dib points out that any marketing plan has one of only three possible outcomes:
  1. Your marketing fails: You make less in profit than you spend on your marketing expenses.
  2. You have no idea if your marketing was a success or failure because you didn't measure the results.
  3. Your marketing succeeds: You make more in profit than you spend on your marketing.
There is a very simple course of action that needs to be followed in each scenario. If you are consistently losing money, STOP and change what you are doing. If you don't measure your results, that's just plain stupid and you need to start tracking them. If your marketing is consistently giving you a positive ROI, then you should GO: throw as much money as you can at it. In the latter scenario, you are printing money at a discount (which is what prompted the $100/$80 question above). Setting a budget is irrational, Dib points out; it implies that your marketing is either not working and is a pure expense OR you have no idea if it's working because you don't measure the results, so you throw money at it in the hope that it's giving you some sort of positive result. It's clear why it irrational to simply not measure your performance, but why does Dib take issue with setting a cap on non-profitable expenses? The answer is because marketing is a profit center rather than a cost center.

Cost Centers & Profit Centers

In most businesses that think strategically about spend, every type of expense is categorized as either a cost center or a profit center. Will spending money on this activity make us money or does it cost us money? The way I like to think about it is in terms of a traditional financial summary workbook: if an expense would typically go on the COGS (Cost of Goods Sold) sheet, it's certainly a profit center. If it would instead go on the SG&A (Selling, General & Administrative) sheet, it may be either a profit center (if it's sales or marketing, for example) or a cost center (e.g. business insurance). We know that Revenue - COGS = Gross Margin. And Gross Margin - SG&A = EBITDA. And EBITDA is, of course, the holy grail: if your EBITDA is healthy then you have a good, solid business; if it's consistently negative you have a business that is losing money and will eventually cease to exist.
So if both types of expenses eventually flow the exact same way into EBITDA, why does the profit/cost center distinction matter? Because it radically changes the mindset of an organization when deciding whether, and how much, money to spend on something. Cost center expenses are to be MINIMIZED. Conversely, profit center expenses may actually need to be MAXIMIZED! Profit center expenses are the $80 you feed into your $100 bill-producing machine. While it would theoretically be even better to spend $75, a winning formula is not something to be messed with lightly.
Another way of looking at the profit/cost center question is by asking: if we don't spend money on something will we, in turn, make measurably less of a profit? The word "measurably" there is really key. Theoretically all expenses of a business help the business function in some way, which eventually results in profits. But if the effect is too far removed from the cause then the thing in question stops being a profit center. Let's say you are in the business of manufacturing widgets and the expense in question are the paper clips used by your home office staff. That's a cost center; unless a cheaper brand of paper clips would somehow significantly effect efficiency in your home office, that expense should be minimized to the extent possible. This is also a simplistic (yet likely relatable) example of why cost center expenses are often viewed as interchangeable commodities—if your favorite brand of paper clip happens to cost twice the normal amount, good luck explaining to your procurement department why you must have that brand. Now let's say instead that the expense in question are the main cogs that power your top line of widgets. Those cogs are COGS 🙂 They are a single step removed from revenue and are, therefore, very clearly profit centers. If you don't buy the cog you don't have a product to sell. OK...but what if we buy a cheaper cog, your procurement team asks? Profit center expenses are much more shielded from commoditization; if these expenses help lend you your competitive advantage, the burden of proof falls heavily on whomever is threatening the golden goose.
So what Dib is really saying is that a marketing expense, which almost all companies view as a profit center, better pay for itself! If it costs you $1,000 in marketing to acquire a single customer, and the lifetime value of that customer to your business is $500, you have a losing marketing plan. You don't budget that type of expense, you change the plan. What if instead marketing was viewed as a cost center, a giant paper clip that gives some essential but immeasurable benefit to the business? Then everything changes, even if the dollar amounts remain identical. The $500 in profit is simply too far disconnected from the $1,000 expense and the two would never be correlated. Setting a budget on marketing expenses in that situation would not just be rational but, like with all "pure cost" expenses, necessary.

BI as a Cost Center

If the title of the article didn't already give it away, anyone who knows me well knew where this was going: business intelligence software and services. Namely, are we paper clips or essential cogs? The sad truth is that, in almost all cases, we are paper clips—a cost center rather than a profit center. A company needs paper clips to organize its documents, it needs Office 365 licenses to enable email and word processing, and it needs BI to enable data analytics. Nobody really stops to ask why they are empowering their users to analyze data. It's just one of those things need to run a business. There's perhaps a general feeling that data must be valuable (after all, we have so much of it!) so if we analyze it enough we can't help but fall backwards into some value. Besides, all our competitors are doing it, so it must be the right thing to do.
Let's say for a moment we accept this premise, that BI is no more than a questionably-necessary piece of general business overhead. Where does that road lead us? Well, given that 90%+ of organizations actually do think that way, where has it led us so far? The first place it had led us is to a lack of accountability. Cost centers are, by their very definition, divorced from revenues and profits. How, then, do you know if you are spending the appropriate amount of money on your BI investment? As a pure cost center, BI is almost always budgeted for; but do these budgets have any meaningful connection to a company's bottom line? Almost universally, the answer is no. Perhaps all the business analysis your employees have undertaken cumulatively has resulted in a $100,000 increase in revenues. Is that a good thing? Sure! Until you realize (which, of course, you never will) that you have spent $1,000,000 on BI software and services. The reason you will never realize how much you have spent on your BI is that the true costs of BI may be so obfuscated that an accurate accounting is next to impossible. Not only may you need to stitch together small "actual" operational expenses from disparate departments, but the true cost in employee time invested in BI is virtually never recorded. Add in consulting hours, infrastructure overhead, cloud computing costs, etc. and you will never get to a real cost number. The other end of the cost-benefit coin is the "benefit" part: very few companies can meaningfully point to a specific analysis effort as leading to a specific increase in corporate profit. So not only will you not know that you have spent $1M on BI, you won't even know that you have earned $100K. The ironic thing is that many enterprises treat BI as a critical business system; an outage of the BI system will be handled as an emergency, with all hands on deck until the system is restored. For all you know, nobody is getting any quantifiable business value out of that "critical" system though. If you are a consumer of BI products or services, that should be deeply upsetting to you.
If you are a provider of those services, such as our own firm, you are likely already upset and frustrated at the commoditization of our industry. As a cost center, BI is, of course, deeply susceptible to commoditization. What makes your brand of paper clips so much better than the dollar-store brand? You have likely had to fight losing battles for the justification of either your preferred products or even your own services. I know I certainly have. You can shout from the rooftops that Qlik (for example) is the strongest software on the market, and can even cite learned treatises to support your position. But if a competitor deeply undercuts Qlik's pricing at scale, your business sponsors are going to feel the heat from corporate finance and procurement to make a switch. The same applies for professional services. You may be able to better help customers feel out requirements, better design a dashboard, develop it quicker, and make it run faster—but we have all seen average billing rates fall as cheaper MSPs have come in with promises of unlimited warm bodies at rock-bottom rates. Can they deliver the same quality as you? Absolutely not (and I think everyone knows it). Does it actually matter though? Maybe least not if we are a cost center. If we are disconnected from profitability, then simply cutting costs is enough to achieve a difference to the corporate bottom line. And the sad thing is that we, as BI professionals, have nobody but ourselves to blame for this commoditization. We have allowed ourselves to become complacent. We don't feel a need to demonstrate the value of what we are delivering. All too often, we take some form of requirements, create some form of BI application, and pat ourselves on the back for a job well done. But is the job well done? Was it even a job worth doing? Have we added value?? And I don't mean in the ephemeral sense of making some analyst's life 5% better; have we added actual monetary value to the company that has engaged our services?

BI as a Profit Center

Something clearly needs to change. I would argue that the current status quo in the world of BI is patently unsustainable. What if...and this is blasphemy...BI was treated instead as a profit center? Now before you all get rocks and attempt to stone me, I want to point out two things. First, you are reading this on a computer screen so you'll just end up breaking your monitor. Second, I have actually seen this done in real life. Admittedly it was only once, about a decade ago, but it made a huge impression on me.
It was a very quick engagement in May of 2012, about half a year before the birth of Infinity Insight. The client was a stock brokerage in downtown Chicago. It all started when my boss called me saying that a new client had reached out to him about an emergency outage of their entire QlikView system. No further details were provided, but I was to go there that very same day to triage the situation. When I arrived, I witnessed a truly hectic scene; papers were lying disheveled everywhere, stock brokers looked beaten down and exhausted. I didn't quite understand what I was looking at yet but it was certainly a dramatic scene, like something out of a movie. I immediately set to work, identifying certain hardware and infrastructure challenges that needed to be addressed to bring the system back to stability. I was about to pack up for the evening and head home when the infrastructure engineer who was assigned to assist me with absolute top priority grabbed some hardware from a shelf and led me in tow (on foot) to a data center several blocks from their offices. He then proceeded to implement all my hardware recommendations that same night. Around 10:00pm, after we had finally stabilized the system, I remarked over a beer that I had never seen such a sense of urgency for a BI outage before. Sure, companies will formally treat outages as emergencies, but the atmosphere I encountered that day was like something I had never seen before. And I had certainly never heard of a corporation fixing hardware issues that quickly. I was used to the calm and predictable world of tickets and escalations; this was like the Wild West by contrast!
The explanation? This firm treated QlikView as a true profit center instead of a cost center. Stock brokers, operating as rapid day traders, had created a system whereby they fed in hypothetical sets of trades into QlikView. The applications were set to reload continuously; the brokers would monitor the results in the application and, if the system of trades they created resulted in a favorable outcome within the span of about the next 60 seconds, would proceed with placing that trade with real money. When I mentioned how much our firm typically charges for last-minute consulting rates, he simply laughed and said: "We lost $15,000,000 today as a result of this outage. The last thing anyone cares about right now is your rate."
I was floored. I had never heard of such a use case before but I knew I was immediately attracted to it. The engagement ended a week later and I went on to have several other interesting and meaningful engagements over the course of the next few months, and founded Infinity Insight in November of 2012. A little while later, I found myself specializing in the pharmaceutical sector —firmly back in the world of BI as a cost center, tickets, and escalations. But over the years, and perhaps colored by that one dramatic experience, I have often wondered what is to stop pharma (or any other industry for that matter) from adopting a BI-for-profit model.

A For-Profit BI Framework

What could a world full of profitable business intelligence look like? It would clearly be very different from the world we see around us today. I don't claim to have all the answers, but a few key components do come to mind:
  • Requisitions: How are new BI applications born today? How are they enhanced? Odds are that the requisitioning department carves out some budget, which they funnel to IT, which implements the application. Or perhaps that department simply makes its case for why an app is needed and the funds come out of IT's own capital budget. In any event, someone has some money that they are authorized to spend and a concrete decision is made to spend it. But in a world where BI is a profit center, and not something that is created "just because" or on the off chance that it may add some value, any requisition of a new application must include a financial component demonstrating the forecasted profitability of the initiative. Budgeting would look quite different than it does today—namely, it would not really exist. Coming full circle: if I were selling $100 bills for $80, how many would you want to buy? If an initiative is reliably forecasted to be profitable, you spend however much it costs to create it. If it isn't, you simply don't do it.
  • Reporting: Forecasted profitability is, of course, not the same as true profitability. And just as a company would not tolerate a profit center where results are simply not measured or one that results in a net loss (going back to Dib's points), careful monitoring is needed to ensure that BI initiatives remain on course to provide a net financial benefit to the company. This would require meticulous tracking and usage systems to be set up for all BI applications, and frequent reporting by accountable stakeholders on financial gains made as a direct result of the BI system. As most forecasts have a time dimension (forecasting quarterly gains, for instance), such reporting would need to occur no less frequently than at this same reporting interval. Detailed tracking and reporting would also provide the company the opportunity to optimize BI systems for increased profitability as time goes on, just as a manufacturer might reasonably expect COGS to go down with experience and scale.
  • Data Quality: There are two important, and distinct, points related to data quality in a for-profit BI application. First, as the saying goes: garbage in, garbage out. Any BI application is only as good as the data that feeds it. When a BI app is directly linked to profits, however, the costs of bad data become much more transparent. As a result, the tolerance for bad or dirty data will likely have to be lowered pretty dramatically. A cost-benefit analysis is likely appropriate to ensure that data cleanliness efforts are robust without overdoing it. Second, there needs to be a serious effort to put the right type of data into our applications. All too often I see disclaimers in corporate applications like "Data herein is not to be used for incentive compensation purposes." If the responsible party cannot load an application with financial-grade data, then what hope does a user have of deriving real profit-inducing insights? There is such a thing as too much CYA, and this is firmly over that line. MapQuest's terms and conditions still state that their maps are for entertainment purposes only; if we don't want our applications to become a joke, then we need to load them with serious data.
  • Commoditization: If we know that an application makes us money, and we know how much money it makes us, that application is much less likely to be treated as a commodity. Even if a cheaper competitor comes along, why would you risk your golden goose for another goose that may or may not be golden? The same is true about the developers of such applications; if a service provider has proven that s/he can keep a key profit-generating business component humming, replacing that person introduces an element of risk that is simply not part of the calculus if BI is a cost center. There are corporate wheels that actively turn against the potential loss of revenue, both in terms of risk and actual loss. If BI falls into the category of "revenue," those wheels will turn in its favor, protecting investments into BI like the corporation would any other revenue stream.
So, in this Edenic world, what happens to all the "other" BI applications that exist in the corporate ecosystem today? Well, to start with, the effects of each one would need to be catalogued and quantified. That's going to involve asking some tough questions, such as:
  • How much do we spend in people-hours to maintain this application each year?
  • How much do we spend on licensing each year? Infrastructure? Cloud compute?
  • How many cumulative hours do employees spend using this application each year?
  • How much cumulative financial value do employees produce with this application each year?
The answers to these questions, which will certainly take time and effort to answer, may destine some applications to the chopping block. Others may need to be dramatically transformed in some way. But while these changes may not be easy to accept, I firmly believe they are necessary if BI is to make a valuable contribution to our clients, and our society, in the future. What do you think?
This entry was posted in General. Bookmark the permalink.

Leave a Reply

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

Notify via email when new comments are added

Blog Home