Beware the Dark Side of Managing with Estimates

Peter Drucker is famous for pointing out how much better managers could manage if they only used numbers. His message to you:“…if you can’t measure it, you can’t manage it.” What often goes unnoticed is that many numbers, estimates are great examples of this, can have unintended consequences in their use.

The primary purpose of estimation is to find an approximate cost for the piece of work prior to it being performed . This can be quite valuable, and the degree to which is accurate it enables us to sell something at a “good enough” price (usually fixed) that would feel valuable to the buyer, and provide enough revenue to you, the seller.

After the contract is signed, though, the incorrect use of the estimate as a managerial tool may actually hurt organizational performance and morale.

Angela walks out of her account review meeting with the CFO feeling like a failure. Her client’s projects are all out of control, showing 5–15% overages every week. She wonders to herself, “how is it that, as hard as I try, I can’t get people to stick to the plan?”

She walks up to Josh, who is at his desk with headphones on and engrossed in his work. He’s a medium-level offender of her project plans, but one of their most talented designers. She and Josh started on the same day a few years ago, and for a long time referred jokingly to each other as “classmates at the asylum.” But lately things have been more strained between them, with laughs few and far between.

“Hey Josh, do you have a few minutes to talk about the Uomo Kitchens project?”

“Uh, oh, oh sure.” He says, setting his headphones down.

She gets right to the point, “I know you’re working hard all of the time, but I just don’t understand why you keep taking longer than planned on this account?” Josh just blinks, so she continues, “Like last week, the revisions to the design was planned at 6 hours and you came in at 8. How could that be?”

Josh looks down at the desk and thinks about how to answer. “I don’t know, really, I tried, but that’s how long it took.”

“I know how long it took.” snaps Angela, “I just had the CFO reminding me of that one and another 10-plus other pieces of work that all went over budget!”

They stare at each other for a second, and Angela continues, “I’m just trying to figure out what I need to do for you to come in on budget, for everyone to come in on budget.”

Josh looks down again, somewhat ashamed, and says, “I don’t know. I’m sorry, but I don’t know what’s going on. I’m doing my best.”

Angela walks away.

It’s a story of betrayal, of hopes and dreams, of human failure, and our confusion over how to manage knowledge workers. Were any of our characters made happy by the exchange? No, not even the CFO. Did this conversation do anything to improve performance? Absolutely not. In fact, Josh is probably performing worse now as the result of Angela’s visit. What’s wrong with this situation? Just about everything.

The above exchange — which is actually quite common — underscores how easy it is to misuse a single number, in this case, the Estimate number versus the Actual hours or cost that is incurred for the work. What Drucker didn’t tell you is that just because you’re using a number, it doesn’t mean that you’re managing well. In fact, it may be quite the opposite.

Estimate versus Actual

So 6 hours estimated and 8 hours actual. Which one is more true, more truthful? The Actual number is factual. It took 8 hours to do the work. Period.

But the 6 hours…where did that come from? It may have come from Angela, it may have come from a template…heck, it may have even come from Josh. But the only truth in that number 6 is this: back when everyone knew less, it seemed like a reasonable number. And that may have been a long time ago.

If we agree that the Actual is Factual, then why do managers compare the Estimate and the Actual as if the Actual is the mistake?

We humans propel ourselves forward through life using hope and expectations of a desired future. If you awoke in the morning and could not imagine a good reason, long-term or short-term, to get out of bed and get going…well, you just wouldn’t move. Ninety-nine-point-nine percent of us choose to get up. Maybe even more than that. We’re pretty much wired to be hopeful. We hope for a 6 every day. Or better.

An Estimate plays right into that: despite being constructed at a point of greater ignorance, we quite naturally hope beyond hope that we’ll get that number right. Sort of like playing Roulette: let’s put our chips on #27 and see what happens on the next spin. The Actual, on the other hand, is the betrayal, the moment when the little ball falls into a slot that is not #27…it is a reason to be upset.

That gambling metaphor may not be all that far-fetched. Let’s look at the 6 hour Estimate and a few of the assumptions:

  • The company had done this type of work for other clients, but not this one, but they saw no warning signs
  • This type of work generally is delivered (the Actual) between 4 and 8 hours, but sometimes as much as 12
  • Angela estimated it two months ago when the contract was signed, using data provided by the project manager
  • At that time, little was known about how challenging the design or the client would be, nor what Josh’s workload would look like

So what are the true odds that two months later this work would take exactly the 6 hours that was estimated? I’m guessing the odds are only somewhat better than our Vegas Roulette bet. And certainly not near good enough to actually get upset when we find out that their estimate was off by two hours. In fact, two hours seems pretty reasonable, given all of the variables that they were dealing with.

Another way to think of this is that an Estimate expressed as a single value demonstrates false precision. because it was a prediction with uncertainty involved, the real estimate was a range, with “6” being point in that range. If the estimate had been stated as “it will likely take 6 hours plus or minus 2” that would be a different story wouldn’t it?

Josh was right: the Actuals just are what they are. Managers make a big mistake when they believe the Estimate, or worse yet, believe it to be more valid than the Actual.

Ironically, the best managerial learning comes from doing the opposite of all of this: looking backwards from the Actual toward the Estimate and the Assumptions. But before we discuss that point, I want to explain a brief history of why we are confused about managing with Estimates versus Actuals.

How the world’s first management consultant led us astray

Almost a century ago, in early days of factory automation and assembly lines, the world’s first management consultant, Frederick Taylor, developed a concept for managing repetitive processes called Scientific Management. Taylor said that managers should measure how quickly workers were doing their work against estimates, or time standards. For example, someone operating a punch press (to punch holes in metal) should be able to punch, say, 250 pieces per hour.

It could be said that time standards were the first “managed” estimates: if we have ten punch-press operators, then how many pieces can we punch in an hour? To be fair, the time standard was actually useful…or mostly so.

Back in those days, the factories were filled with simple machines and the workers were usually uneducated farm workers or immigrants who had just arrived in the US. Measuring how well people stacked up to a time standard actually made some sense for this type of boring, repetitive work, More so, if we assume it was being done by people who probably didn’t have a passion for what they were doing.

Taylor was famous for saying that workers would work at the minimum speed that goes unpunished. And that one idea, that workers are intrinsically lazy, became part of our social fabric. Further, he said, it was the job of management to enforce worker performance through threats and punishment. Wow.

Within a decade or so, Taylor admitted that the whole approach was somewhat “dehumanizing” and that there were probably better ways to manage. But we seem to have never shaken his words out of our heads.

In our vignette, the CFO was speaking to Angela using much the same logic: why weren’t the workers meeting the numbers?

In Taylor’s world almost all work was repetitive and should yield better performance through management enforcement. But by the 1960’s the Japanese had shown that uneducated workers performed even better on repetitive tasks when allowed to manage their process and the production line. Implicit in the approach that the Japanese took, was a fundamental belief in our the goodness of our humanity: that people do want to perform well (like Josh did) and probably care as much or more than the managers do.

The Japanese created a revolution in factory management by doing the opposite of what Taylor suggested: they started asking why the Estimate was wrong. It turns out that in almost every case, not meeting the Estimate was not about people’s limitations, failures or laziness, but rather, about broken processes and bad operational assumptions.

Why didn’t the age of knowledge-work obsolete Taylorism?

These days, truly repetitive work is done almost solely by machine. Most people like Angela and Josh are knowledge-workers: their work is complex and subject to variability or uncertainty, enough so that a human is required to adapt, improvise, ideate and execute flexibly.

This is not Taylor’s world. There is no time standard for work that is non-repetitive, like knowledge work. Why would we assume that the model that didn’t work all that well in manufacturing would work better in knowledge work?

Apparently, nobody told the project management and time-tracking software companies…and maybe even many CFOs. When you see a system or a spreadsheet that displays “variances” and when the Actual is larger than the Estimate number, which one is more likely to be colored RED?

Yes, the Actual is either displayed as the problem or discussed as if it is one. Ironically, by doing this, we also ignore something else: estimates that are too high. If the Actual comes in below the Estimate, we also estimated poorly. So why isn’t that number red also?

Want to manage projects better? The simple and non-intuitive managerial shift is to focus on why the Estimate was wrong. The correct use of the Actual is to help you create a better Estimate and diagnose your assumptions and processes. Once you start the habit of doing that, you’ll create better conversations.

Angela and Josh are sitting down to discuss the latest variance report.

“Wow,” says Angela, “we keep getting these estimates wrong! Do you have any idea what we’re missing?”

Josh pauses thoughtfully and then says, “I think we’re not factoring in that every time we think it is going to be simple, the client throws us a curveball. And also, they keep on changing priorities. I lose work time just changing from one job to another.”

Angela smiles and says, “Yeah, that’s like a change order, but we don’t really count it that way, do we? I wonder if I can change the way we bill this client so that we encourage them to stay on-point with priorities?”

“And also what could be helpful…” Josh looks at Angela to see if she wants to hear more. Angela nods and he proceeds, “what if you and I only talk about the exact work that I’m working on, not other topics? Sometimes you come by after seeing the client and you are so excited. I love hearing about the new work, but it also is disruptive to my day. Could we sync up on your news, and mine, at the start or end of the day, so I can be more productive?”

“Totally. I had no idea that would be helpful to you, and yeah, I do get excited and want to share it with you. You’re my classmate!” Angela grins and Josh smiles back. Angela adds, “I’ll go let the CFO know what our plan is for getting those numbers to come closer to each other. Thanks for your help!”

In the end, estimation is still a useful and valid exercise. But you need to be cautious as to what the number means and how it is used. Here are a few tips from my playbook:

  • Scope twice, execute once. Often we sell work with an optimism that is unwarranted in delivery. Call it what you want to get the deal. But make sure you scope again once the deal is booked, or after time has passed, or you know more.
  • Teams estimate better. A large body of research (and my experience as well) suggests that the more senior you are, the more optimistically you estimate. Delivery scoping and estimation is best done with teams.
  • Document your estimates. I don’t mean an “assumptions” section that your client neither reads nor understands. Each piece of the work needs to have its attributes spelled out — both what is included, and what is not; in this way, you can think of “negative scope” as being the most important part of scoping.
  • Consider context. After you have thought through what the work is, take a moment to think about the other parts of getting it done, like the environment that people are working in (how interruptive?), how much time the team will spend having to coordinate with each other, how many reviews the client will get, or even how well understood the work is in the first place. In our work with complex organizations (I call them NOCOs) these contextual and process factors can have a greater than 50% impact on the gap between your Estimates and Actuals.

Taiichi Ohno, the architect of Japan’s manufacturing revolution, always said that failure was a better teacher than success. By changing the estimates-versus-actuals conversation on every level, you can let the Actuals be your teacher, guiding you to ever-better performance.

I’m also hoping that in shifting your thinking slightly, you become a bit more skeptical (or at least curious) about how to correctly use other numbers when managing.

Don’t let me get started on timecards.

Jack is a recovering client services executive, and the founder and CEO of AgencyAgile, a productivity training and coaching firm that helps agencies, marketers, and other complex service organizations.