The case against #NoEstimates: the bottom line

I’ve now methodically presented the case against #NoEstimates in three different lights: from a common sense standpoint, from the perspective of the solid reasons why estimates are useful, and by examining the various frequent talking points used by NoEstimates advocates.  Looked at from any of these angles, NoEstimates comes up way short on both its core ideas and business practicality.

Aside from these issues of substance, let’s look briefly at the behavior of the NoEstimates proponents. Blunt as it may be, here’s my summary of the behaviors I’ve seen across most NoEstimates posts and tweets:

  • Presenting, and repeating via redundant tweets month after month, fallacy-riddled arguments consisting primarily of anecdotal horror stories, jibes at evil management, snide cartoons, and vague declarations that “there are better ways.”
  • Providing little or no detail or concrete proposals on their approach; relying (for literally years now) on stating that “we’re just exploring” or “there are better ways”
  • Consistently dodging substantive engagement with critics, and at times openly questioning whether critics should even have a voice in the discussion. If NoEstimates avoids engaging actively in the marketplace of ideas and debate, why should their arguments be taken seriously? Real progress in understanding any controversial topic requires we do more than state and restate our own views, but actually engage with those who disagree.
  • Continuing to use discredited examples and statistics, or even blatant misrepresentation of the stated views of recognized authorities, to help “prove” their case.
  • Frequent use of epithets to describe NoEstimates critics: “trolls”, liars, “morons”, “box of rocks”, and more.

I pointed out in my introduction that the lofty claims of the NoEstimates movement (essentially, that software development can and should be an exception to the natural, useful, and pervasive use of estimates in every other walk of life) carry a heavy burden of proof. Not only have they failed to meet that burden, they’ve barely attempted to, at least not the way that most people normally set about justifying a specific stance on anything.

But aside from style, let’s return to the substance of the issue. Here’s my take, as backed by specific examples over the course of these blog posts: estimates are an important part of the process of collaboratively setting reasonable targets, goals, commitments. Indeed, whether estimates are explicit or implicit, they’re a reality. I see them as an unavoidable and indispensable factor in business.

Fundamental truth: successful businesses (and successful individuals) achieve their success by setting concrete, measurable goals for themselves, and then stepping up to those goals. Because goals matter, and when a needed result will occur matters. Hundreds of related business decisions and actions usually depend on the “when”. Companies on Wall Street, if they want to get any coverage by analysts, typically set guidance (in essence: predicting the future) on expected business results. Why does a company’s stock tend to crater if the expected business results aren’t achieved? Because investors know that a failure to hit targets tends to correlate highly with being a business that simply won’t succeed over time. Such a failure signals investors that the individuals in that business, for whatever reasons, good or bad, internal or external, just aren’t getting it done.

#NoEstimates embodies the lamentable and intentional antithesis of such healthy goal setting. Aside from the fact that their approach (to the minimal degree they’ve outlined their approach at all) eliminates most of the direct benefits of overall project goal setting and tracking, NoEstimates advocates seem particularly unaware of the business perception toward abandoning such goal setting: specifically, they ignore how people in the business react to the NoEstimates extreme insistence that things in software are just so uncertain, how you just can’t predict the future, and how estimates are abusive.

This sort of uncertainty-resistant mindset isn’t specific to NoEstimates, of course, but can run rampant in IT in general, as I’ve discussed. If asked to set achievable, concrete goals, many IT people will go on and on about all the times they’ve seen goals set for them that weren’t achievable. If the topic is making solid, reasonable commitments, many IT people will turn the conversation to all the times they’ve had unreasonable commitments dictated to them. Emotions and self-righteousness run high. Management is often depicted as completely clueless, if not totally evil. Check out the embittered tone of most Dilbert strips for telling examples: amazingly, one major NoEstimates advocate posted a link to the same Dilbert strip six times in a single day and dozens of times in a month.

These, especially combined with the pervasive NoEstimates rancor towards critics, are veritable PTSD-like behaviors, frankly; they should be regarded with an appropriate level of empathy for the stricken, but let’s work hard not to have these damaging symptoms spread to the unafflicted.

Don’t get me wrong, though: all of those counterexamples, those examples of dysfunction via the misuse of estimates, need to be taken to heart, and then assiduously managed around so they don’t happen. But a very quick way to lose respect in a business environment is trot them out incessantly and then have people conclude that you’re dodging judicious, collaborative goal-setting and commitment. There’s not a chance you’ll get that long-desired seat at the table then.

Again to harken back to Pogo: we’re our own worst enemy.

We seem to think up all kinds of reasons not to play in the world of business realities. And by doing so, we marginalize ourselves; we become the geeks in the corner, the mocked members of the school’s A/V club, the people who aren’t expected (or welcome) to discuss anything of true meaning to the business. Service providers. Order takers. Seen (and resented) by many as little more than reluctant and ornery mercenaries.

Professionals, on the other hand, in any field, don’t chronically complain that something has never been done before and simply can’t be predicted. They use their best professional experience and trained judgment to evaluate what it will take them to accomplish the given task, and they’re willing to take a stance on what that necessary level of effort could be. They don’t grumble that someone asking them to give such assessments is “knocking the ice cream out of their hand“. They collaboratively establish reasonable goals with their customer, they stick to them within reason (absent, of course, new, game-changing information), and they tend to deliver on them more often than not.

In IT, as in life, it’s wise to be wary of the extreme, the obsessive, the “one note” tirades. The NoEstimates arguments, as I’ve shown, are ill-considered and poorly argued, and they unfortunately seem to appeal disproportionately to the inexperienced and the disgruntled. If no one speaks up, then even bad ideas can propagate and cause real damage, especially when evangelized so persistently. So it’s important to push back on the NoEstimates wild claims, the misinformation, the misconceptions, because that’s indeed what they are.

Estimates aren’t guesses, and if someone starts to tell you they are, look for deeper underlying motivations such as chronic disgruntlement with management or an obsessive desire not to be blamed for anything. Estimates are also not plans, guarantees, commitments, or a schedule. Estimates are tools, common and pervasive and useful. And as tools, estimates matter in business, to anyone who is thinking beyond his or her immediate silo.

The answer should be obvious: rather than abandon or avoid estimates, get good at them instead. (Here’s a great list of starting points to learn how to do estimates more effectively.) Fix your processes if your estimates are chronically off and that’s hampering your overall ability to serve the business reliably. Earn the trust of your customers by collaborating with them on setting viable goals and schedules, and by delivering according to the evolved plan.

And that’s the bottom line.


  1. Extraordinary claims require extraordinary evidence, – Carl Sagan.
    There is none forthcoming.
    The claim is …
    ◆ There are decision making framework for projects that do not require estimates – likely domain dependent. No stated domain constraints. In the absence of a domain and constraints on the application of estimates, it’s hard to tell.
    ◆ There are investment models for software models that do not require estimates – investment models are about future returns on the investment, and must deal with the uncertainties about this future. Assessing these uncertainties requires estimating. This is the basis of all investment models.
    ◆ Project management (risk management, scope management, progress reporting) approaches that do not require estimates. Risk management is about the reducible and irreducible uncertainties that create risk. These uncertainties – aleatory and epistemic – require estimates of their values. Aleatory uncertainties are assessed from the Probability Distribution Functions and the random variables produces from that PDF. Epistemic uncertainties are assessed from the discrete “probability of occurrence,” and modeled from the estimated probability of occurrence and probability of impact.

    These statements have no logical foundation in the decision making paradigm where the future is involved – IF that future is not known in a definitive manner – that is there is no uncertainty.

    In the absence of actual explanations, these statements are simply unsubstantiated conjectures. We’ll have to wait for the book to see if there is any substantiation that can be tested outside of the personal opinion.

    In other words – Show Me the Number.


  1. […] Lean Agile Manchester but there’s still a lot of debate and discussion on various blogs. This post by Peter Kretzman caught my eye and is well worth a good […]

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.