My Analytical Story
I have loved business my whole life. When I was a little boy, my grandma would take me down to the local toy store and I would buy Monopoly money (you didn't have to buy the board game, you could just buy extra money)...
Then, I would put on the suit my parents bought for me and walk around with the monopoly money in my pockets feeling like I was a businessman. The first business I was part of was selling corn on the side of the road with my Aunt Coreen. It was just so amazing - we would have to decide the price of the corn, make signs, pick the corn, find the perfect spot on the road to sell it and work hard all afternoon servicing our customers. I loved every part of it, and I can't wait to run a small business like this with my kids. My second business was in Junior high school. My Dad would travel to the US for business and buy a case of peanut M&Ms at the US Costco. I would buy them off my dad for $0.30 each and sell them for $0.50 a pack. It was tons of fun and I would sell out my supply in a few days and then have to wait until my Dad went down to the US again. In grade 10 I read my first business book, "What they don't teach you at Harvard Business School", which kicked off my love of reading. I went to University to major in marketing but quickly also fell in love with economics, management accounting, financial analysis and statistics.
My second job after University was at a small financial institution in Alberta where I was in the marketing department. It took me a while but eventually I realized a great business is founded on insight, and insight is best built off analysis, which means mastering accounting, finance, statistics and consumer insights. This path eventually let me to the conclusion that the very best insights come from fusing multiple disciplines together. The first campaign I had to put together was for mortgage season. Every bank, including ours, would advertise mortgages using mass-media and targeted direct mail to existing customers. To kick off my thinking of how to do the campaign I went down and talked to a few mortgage managers and asked "how do consumers decide where to get a mortgage?". One manager's response was the nugget of insight I was looking for - that all our customers look around for the best rate and then use that as negotiating power to chew us down on our rate and it was almost our decision to match that rate or let the customer go. All our competitor's customers were doing the same, phoning us to find a better rate but their current bank would often match it and keep the business. So the key to getting a mortgage was not building a marketing campaign around a great rate unless we had a fundamentally different business model that had a lower cost structure to allow us to quote a rate the competitors won't match. The key had to be something different - you had to get them as a customer first, then you had the chance to win their mortgage. So how do consumers decide which bank is their main bank? Whose customer is whose? That I had to research.
So I went back to my office and dove into our CRM and what I found formed the initial foundation of what today Fusion calls the Revenue Opportunity Chain: that the most important way to segment a customer is not by demographics, or life-stage, or psychographics, but by their orientation to the business. We needed to determine whether they were an existing core customer, marginal customer or non-customer, and this classification would drive our marketing, messaging and pricing. In our trade area (where we had branches) approximately 10% of the population had a chequing account with us, and during a mortgage campaign we would advertise to the whole population. However, 95% of our business would come from the 10% of the population that already had a chequing account with us. The same for RRSPs. The same for any campaign. And the real kicker was we had no strategy actually aimed at getting more chequing accounts, even though that was by far the determinate factor driving the business' success. This insight was transformational; the way to get more mortgages wasn't to advertise mortgages more but to advertise chequing accounts more. The way to get mortgages wasn't through a lower interest rate campaign (though we did need to aggressively match rates) but through more tellers, higher chequing overdrafts, more ATMs throughout town. If you remember back to the 90s all the retail banks were cutting hours, having branches with no tellers or even closing branches. But what were we doing? Adding hours, adding tellers, adding branches and improving our chequing services. And all of this success started with insight, it started with me walking down to the branch floor and listening to the front line staff and then fusing this insight with data-mining, market research and financial analysis to uncover something nobody else knew in retail banking 15 years ago.
Years later when we were proven right in our strategy I started to wonder "why did we uncover this and not the banks? Why was their strategy so flawed when they had access to way more information, analysis, insight and consultants?" What I began to realize is most decisions are first made not with insight, but with an idea. The idea could be the "internet is the next big thing" or "ethnic consumers are growing" or "we need a loyalty program." Usually these ideas are smart and hard to argue with. Everyone knows the internet is the future. It is what Nobel Prize winner Paul Krugman calls "what the Very Serious People know." All the Very Serious People know the internet is the future, and if you don't understand this you are not Very Serious. And they are right, but in important ways they are also wrong. This approach of coming up with an idea and then testing if it is true, what you could call developing a hypothesis and then doing analytics to validate it, is extremely blinding, for often something can be true but still not be the best path. Today our firm Fusion Analytics calls this "bottom-up thinking" vs. the better approach of "top-down thinking." When faced with a problem or decision, we believe the key is to start the conversation not with "how did X do?", or "how do we do X better?" or even "should we do X?". At Fusion we believe that instead of starting with the "X", the resource, we start with the issue, asking "what is the best way in the next 12 months to grow market share?", and working our way down. This sounds simple but it is transformational. I believe it is the main reason we uncovered the right way to do retail banking, and is the main reason Fusion is able to bring new thinking to the table and really take retail analytics to a whole new level.
After the financial institution I went for my MBA in London Ontario at Ivey. During my summer internship at a major fast food chain, I learned another valuable lesson: that business is about making the right decisions, and decisions are based on how we view the world, the results we see from the actions we have taken in the past, how we size up competitors, how we see consumers and how we understand our financials. All of this means we need the right data, the right analysis, the right metrics and ultimately the right questions. Nobody at the company was spending any time worrying about collecting the right data, data quality, building better analytical formulas, understanding how the metrics worked or if we had the right metrics at all. One metric that was beyond the pale was the percent mix of sales of a new product. This was basically the only metric the company used to measure a new product's success: the higher the mix the better. But here is the rub - almost all new products had a lower profitability than the existing mix, so if the new product was just cannibalizing other products then the company was worse off, but if the new product sales were incremental sales then it was winning. So, my view at the time was why wasn't the metric incremental sales instead of product mix?
After thinking about this for eight years I have decided the best answer to why these horrible metrics won't die is because most people shy away from ambiguity. People prefer a hard, cold precise number over a fuzzy approximation, even in situations where a fuzzy number like incremental sales is clearly more accurate (i.e. will lead to a better insight and action) than using the more precise but at best useless, and at worst counterproductive, metrics like sales mix and flyer sales.
When I started to notice these poor metrics I looked around and what I saw got me extremely depressed. We live in a sea of horrible metrics. Retailers ask their customers if they were satisfied with their last purchase but (and this is a very large but) we only ask it to those that were satisfied enough to make a purchase. We are literally only asking satisfied people if they are satisfied. For flyer performance we measure flyer sales, or flyer sales as a percent of total sales or flyer sales per square inch. But if all you are doing is putting something on sale that you would have sold anyway, and you're not driving incremental sales, then you are not helping your top line and you are killing your bottom line. Despite these inherent flaws, it is almost impossible to get rid of these horrible metrics. But metrics are the fundamental way we learn at companies; they form the information that builds decisions. In my view nothing is more important than fixing them. My mission in life, the mission of Fusion, is to bring better questions, metrics, formulas and data to retailers. When I started Fusion six years ago I had two concepts in mind:
- That retailers would benefit from better metrics across all disciplines, from marketing to merchandising to pricing, real estate, operations, inventory, etc.
- That retailers have a need for historical tracking, but there is a blind spot regarding the equivalent need for forward-looking opportunity modeling. Regardless of the discipline, be it price elasticity tests, econometric-based modeling for media mix or run-rates over LY for inventory demand forecasting, almost all analysis is focused around just historical measures, which is important but just half of the puzzle. There is a big win to adding opportunity modeling to the mix.
It also seems to me that over the last six years the retailers that are growing dramatically are the ones who grab onto this kind of thinking and embrace it. The future of retail analytics is bright, if retailers have the knowledge and motivation to ask the right questions and seek the right insights. The good news is we are starting to make a difference; we now have a dozen retailers working with Fusion on a whole new and better way to use information to make decisions. One of our clients calls it "taking the red pill." In the movie "The Matrix", the main character Neo is offered the choice between a red pill and a blue pill. The blue pill would allow him to remain in the fabricated reality of the Matrix. The red pill would lead to his escape from the Matrix and into the real world. For Fusion the red pill means beginning to think less about tracking and more about opportunity modeling, less about bottom-up thinking and more about top-down analysis and worrying less about being precise and more about being accurate. For 2014 we hope even more retailers take the red pill.