Source: Scott Adams
When I started my first job as a management consultant in the late 90s, my manager had a Dilbert cartoon posted over her desk, entitled “How Estimates Become Facts Through the Miracle of Communication.”
It depicted Dilbert’s infamous Pointy-Haired Boss asking him how big a certain market was.
Dilbert turned to his hapless colleague Wally for help.
“I don’t know,” Wally responded. “It could be anywhere from one to a million.”
“It could be a million,” Dilbert wrote in his report.
“Experts say one million,” the Boss told his board in the final panel.
The cautionary tale was particularly important at The Boston Consulting Group. No sooner might we share our own qualified estimates — typically a blend of Internet- or client-sourced factoids and “BCG analysis” — in a client or public report that the same would begin to appear elsewhere as an unqualified fact with the grand label, “Source: BCG”.
Much of the business world operate with more estimates than facts. Even something as apparently objective as financial statements are informed by necessarily informed by guesses, on everything from the market value of inventory to the depreciation of assets.
This is not a bad thing. Just a necessary practice in a complex world that must be reduced to something more manageable on paper for decision making purposes.
We are all seeking to work our way through a non-linear reality, full of countless moving parts, with our linear brains. While data gathering and tracking has come a long way since the last century, we still need to make choices far more quickly than the time it would take to get a full picture of the data we require to make them.
There’s no pause button. By the time we could ever get the full picture of reality we might desire, the facts on the ground will necessarily have changed. To see this in practice, look no further than the business pages, where central banks the world the world over are constantly challenged to make rate decisions based on macroeonomic information that’s at best a quarter or two out of date.
And, of course, full of estimates.
What’s a manager to do?
The short answer, as most will quickly admit, is not much. This is simply the way life works. We are barrelling down the highway in vehicles of various sizes — from Fiat-sized startups to eighteen wheel-public companies — with much of our windscreen blocked and dashboard instruments telling us how fast we were going or how much we had left in the tank an hour ago.
Yet our most successful leaders get that there’s a bit more to the picture, and engage in three critical practices.
1. Call B.S. On Your Sources.
“Your father did business with Hyman Roth,” the doomed Frank Pentagelli told Michael Corleone at the beginning of Godfather 2. “Your father respected Hyman Roth. But your father never trusted Hyman Roth.”
Neither, as we gradually learned, did Michael. And the manner in which he sought to work with such a treacherous partner provides a model for how we should deal with information sources we might find anywhere — be that the Internet, “expert” reports, even consultants or employees who feel the need to come up with some kind of number.
For simplicity let’s zero in simply on the first of these. If you’re reading this post, chances are you’re sophisticated enough not to believe everything you read just because it popped up in a Google search.
You need to be confident that anyone who might supply you with numbers feels the same way.
The biggest risk is around estimates that have become facts via something like the Dilbert model above — simply becoming ubiquitous.
Each of these claims has a basis in reality to the extent that someone said them sometime in some context.
Dig a little, however, and it quickly becomes apparent that they are actually meaningless statements.
Jo Cook of Light Bulb Moment has posted an excellent post thoroughly debunking the latter. Not only did the original study put forward the hypothesis — not a claim — without data and with caveats. Most Internet references to this fact have referenced each other without every reading the original study. Indeed, Cook even found the claim being attributed to her own article that questioned it.
The solution is not necessarily to engage in such detailed investigation ourselves — although we’re often only a handful of clicks away from the original source, and it may take only a few minutes to read for ourselves.
That takes time that most managers don’t have. A simpler approach is simply to ask ourselves whether such claims, no matter how many times we’ve heard them, actually make sense.
Imagine yourself making a pitch to a new client, or a strategic recommendation to a board, without speaking. Could you get 90% of your message across?
“Ninety percent of what?” Franklin Jonath, a renowned psychologist and communications expert, once told a group of trainees. “That you’re angry, or delighted? Sure. That we should grow in Europe rather than South America? Good luck.”Likewise, try making a presentation about how your revenue mix is evolving without visualizations, such as a value chain, pie charts, or maps.
2. Structure (And Honor) Your Guesstimates.
“Lean” does not mean “cheap”, as the name might suggest. Rather it refers to the elimination of waste — specifically, by approaching each new initiative not by going on all in, as though it simply has to work, but as simply a best guess that must be tested and de-risked.
The Lean ‘build/measure/learn’ mantra has become ubiquitous, even an article of faith, particularly in tech. Throw a stone into any crowd of entrepreneurs, and you’re guaranteed to hit someone who can rattle off the theory of how it’s supposed to work.
Finding a team that actually practices it as its authors intended is much harder.
Why? Because it’s not just about building whatever you like, willy nilly, and seeing what happens. Outside of the most advanced data shops with large teams of analysts, you can’t just measure everything and see what if anything jumps out.
The hardest part of the Lean loop is “measure”.
The starting point of any successful learning loop lies in structuring a sense of sensible, testable hypotheses. Even the simplest product experience needs to be broken down into stages, with appropriate measures in place, and your best guess as to what you might learn from higher or lower numbers.
This means — yes — making estimates first. You need to think hard about what you know, kind of know and don’t know before you make a change to a process or write a single line of code.
Guessing well is harder than throwing stuff at the wall and seeing what sticks. As any capable engineer or management consultant will tell you, guesstimation is not the intellectually-lazy exercise the word might suggest. It’s a systematic practice, even a science, that relies on proven techniquesand gets better with experience.
In some businesses, such as commercial construction, estimators are the main load bearing element, whose capability represents difference between success or failure — like the goalie of a hockey or soccer team, as a client once pointed out to me.
Estimating is not a secondary competency. It’s not a backup to the “real” stuff. It is, for all businesses, core — and must be honored as such.
3. Treat Everything As Research.
You don’t need to wade through hundreds of pages of research (though I highly recommend it) to grasp the fundamental sense of this, and what it means for all business.
Profit is the difference between what someone will pay for a good or service (its value) and how much it costs to produce and market.
The more a company understands its market and its processes, the greater this profit will be. Which in turn creates dividends for investors that can either be spent on other goods and services (generating more profit in the system) or reinvested in developing new goods and services.
Some products will fail. Some businesses will fail. But the trend of human economics is that the learning from such experiences gets ploughed into developing others that will win — and so continue what, to date, proven to be an inexorable and astonishing upward rise in human wealth.
I often hear clients refer to the necessity to “do research” before undertaking a particular initiative, be it developing a new product or entering a new market.
In many cases this can be an invaluable activity. Rather than assuming a given geography is favourable for new industrial development, for example, a business can learn a lot about its prospects for success (and the required ingredients) by reviewing publicly available data and conducting private inquiries of its own (which is where many of us management consultants earn our keep).
But — and this may be the most critical point of all here — such research must be treated with due caution. Most of it consists of the estimates of others — and no matter how impressive the brand or the letterhead, it should not be mistaken for hard knowledge.
Indeed, I encourage my clients to think of “research” as little more than a way to make sure we structure our own learning as efficiently as possible. And what is “our own learning”? It’s getting out there and trying to build something, be it a partnership, a customer relationship, or literally, a building.
Steve Jobs was often asked about Apple’s approach to product development, particularly after the breakout success of the first iPhone in 2007. His biographer Walter Isaacson famously quoted him as saying:
People don’t know what they want until you show it to them. That’s why I never rely on market research. Our task is to read things that are not yet on the page
This did not mean that market research had no place in Apple’s arsenal, as it clearly does. Simply that the last mile of any guessing or estimation exercise comes at the coalface — actually putting out a product.
No matter how sophisticated data and analytics tools become, managers will always be working, necessarily, in a state of ignorance. The challenge is not to change this natural law, as immutable as gravity. The challenge is to accept this as a core reality — and proceed accordingly.