BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage News Planning with #NoEstimates

Planning with #NoEstimates

Bookmarks

People are used to making plans and taking decisions using estimates, even though they are often not so good in estimation says Gil Zilberfeld. The #NoEstimates movement explores alternatives for estimation. At the Agile Testing Days 2015 Zilberfeld will do a workshop on Planning with #NoEstimates.

The Agile Testing Days will be held from November 9-12, 2015, in Potsdam Germany. The motto for the 7th edition of this conference is:

Embracing Agile for a Competitive Edge - Establish leadership by delivering early, rapid & iterative application releases

InfoQ will be covering this conference with write-ups, Q&As and articles.

InfoQ interviewed Zilberfeld about why people do estimation and how they look for ways to get better in it, and had a discussion with him about alternatives for estimation.

InfoQ: Can you briefly explain what you mean by "Planning with #NoEstimates"?

Zilberfeld: We’re so used to base our planning on cost estimates. It’s so ingrained in us. When you have an idea, or a solution, the gut reaction of management is to ask "how long will it take" or "how much will it cost". And then we decide to do it based on the answer.

That’s a shame, because we humans are really bad at estimating. We should make decisions based on better methods, and there are alternatives. The workshop is based on alternatives to estimates, that you can find around the discussion around the #NoEstimates hashtag.

InfoQ: Why do people do estimation, what do they expect from estimates?

Zilberfeld: It’s very interesting. If you ask people, they will tell you that without an estimates they can’t make a decision about a project.It’s like they are stuck until you give them a number. The thing is that we don’t really *need* it to make the decision, but we *want* it. We want to make confident decisions, and when we get the number we try to see if it fits our experience, point of view, ideas and knowledge. If it fits somehow, we can make a "go" decision with confidence.

The other part is feeling in control. The estimate is like a fog vacuum - it clears all the uncertainty away. If that number fits with our views, it gives us confidence we can ride the tiger - control the project - when things go wrong. All it takes is the right number :)

InfoQ: It appears to me that most people are not good in estimation. Are there ways to get better in it?

Zilberfeld: Sure you can get better at estimation. The problem is that as good as you get, reality will bite you. There’s a big disconnect between how we think things will go and what will actually happen. We’re optimistic, we think we can control the tiger, and our view skews our numbers.

Hofstadter’s Law says: It will always take longer, even if you take into account Hofstadter’s Law. We play a game in the workshop that gives all the conditions to get estimations right and almost all people get them wrong.

We should remember that we use estimates to make decisions. That’s only one way, and if we are not good at it, maybe we should try other methods.

InfoQ: If estimation is not the way to go, what’s the alternative that you suggest?

Zilberfeld: We can look at the team’s track record - if they did something similar in the past, chances are you don’t need the estimate, but use the old actual cost. If however, the team never did this before (or what if no one did this before?), there’s not much point in asking for a number. What good will it do? Instead of making a go/no go decision around a 2 year cost estimate, we can budget small experiments that will tell us if we’re on the way to achieve the grand goal. We change the way we’re planning, from big up-front plan, to deliberate discovery using cheap experiments, which we can stop any time.

Finally, if we’re so keen on estimating, why not estimate business value, rather than cost? These are still estimates, we can prioritize projects based on value, rather just by their cost. When leaving out business value, we make really bad decisions.

InfoQ: Does #NoEstimates mean that you should never ever do any estimates? Are estimates evil?

Zilberfeld: #NoEstimates sounds like it’s about abolishing estimates. But it’s really about searching for alternatives. The question comes up frequently, but the answer is the same from all the #NoEstimates proponents - if asked for an estimate, give an estimate. Estimates are not evil, but they are very much abused. They are often thought of as promises rather than guesses. So when the estimator misses, there’s disappointment and sometimes anger. And maybe worse, when the estimator succeeds, his expert authority goes up, discarding any uncertainty that was in the process.

But that’s not all. Since people miss estimates all the time, we see managers "extending" estimates given to them. If I say something will take 3 weeks, my manager will try to buffer it saying it will take 5 weeks, and her manager will make sure we don’t miss by saying it will take 8 weeks. So a short project becomes a medium one, with all the committed resources for two months. As with the Student Syndrome, work expands to the time allotted, and that’s quite a waste.

Finally, if estimate don’t provide us with the tools we need, the whole planning process becomes wasteful. And if cost estimates are the only decision making tool we have, we’re going the way of the gut-feel, rather than the data-driven road. When you understand what you really get from estimates, it really doesn’t make sense using them exclusively anymore.

InfoQ: If people want to learn more about planning with #NoEstimates, are there some resources that you can recommend?

Zilberfeld: Apart from my workshop, you mean? If you want to learn more about #NoEstimates follow Woody Zuill, Neil Killick and Vasco Duarte on Twitter and join the conversation. Vasco also has a new book called, surprisingly enough "No Estimates". More and more people are now contributing stories how they manage with alternatives for estimations, and maybe you can pick an idea to try with your team!

Rate this Article

Adoption
Style

Hello stranger!

You need to Register an InfoQ account or or login to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Community comments

  • Cost vs. Business Value

    by Fredrik Vestin,

    Your message is awaiting moderation. Thank you for participating in the discussion.

    "Finally, if we’re so keen on estimating, why not estimate business value, rather than cost? These are still estimates, we can prioritize projects based on value, rather just by their cost. When leaving out business value, we make really bad decisions"

    Problem is that if it's hard to estimate time (i.e. cost) it is even harder to estimate business value. Typically you have no idea until he stuff is built, released and you can gather feedback and data. Up until then you are guessing.

  • Re: Cost vs. Business Value

    by Gil Zilberfeld,

    Your message is awaiting moderation. Thank you for participating in the discussion.

    Nobody said we have an easy job :)

    Estimating value, using Cost of Delay for example, is building an assumption set that at least can be debated. You can agree or not with my assumptions. If at least we understand the assumptions we can base our decision more than a gut feeling. It's much better than make decisions based on cost guesses.

    If you don't know what will happen (like most people) you'd better change the way you work - instead of high risk big projects, to low risk experiments that will help you learn about your assumptions. As you learn, you'll make better decisions, hopefully at a lower tuition cost.

  • Still estimates

    by Henrik Ebbeskog,

    Your message is awaiting moderation. Thank you for participating in the discussion.

    Hi!

    Here's my view on the "alternative". You say 2 years cost estimate. Who says it's 2 years (or whatever)? And why do someone think that? This is estimating. And if you have no idea - at all - if it's 2 weeks, 2 months or 2 years, you'd better get this idea. Otherwise there are possible opportunity cost losses here.
    And even if you start anyway and do "small experiments" (which I fully support) you might end up at 2 years before grand goal is reached.
    And still, "small experiments" require estimates as well. Like "How small?" "What and where is a good validation point?" etc.

    There's always estimates. No way around it.

  • Re: Cost vs. Business Value

    by Fredrik Vestin,

    Your message is awaiting moderation. Thank you for participating in the discussion.

    Sorry for the late follow-up. Cost of delay assumes you can guess the business value of a feature, again this is very difficult and definitely and in my opinion no easier than guessing the time it will take to develop a feature. The alternatives that you present are also estimates. Looking at the track-record of the team and assuming that we can get the same amount of stuff delivered in the coming sprint is also an estimate. Other #noestimates suggestions like splitting stories until they are all the same size and then counting finished stories is also estimating. In my opinion estimates are not evil, just don't abuse them and think they are guarantees and waste too much time estimating.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

BT