If I had a tractor…

Agile Estimation and Planning

Introduction

With the advent of agile methods and their affinity for rapid delivery cycles, traditional estimation models are coming up wanting.  Agile estimation, while undoubtedly faster is also difficult for some to grasp.  Expressing work in terms of time has been burned into the software community’s psyche.  But it doesn’t have to be that way.  I coach agile teams in making the switch from traditional development models to agile.  In doing that, I help them learn new ways of estimating.  This series of articles will hopefully explain – through metaphor, why the old way of estimation was problematic, and how the new way addresses those shortcomings, is faster, is no less accurate, and can even be fun!

If I were to stand up in front of a room full of people, and select individuals at random to answer this question:  “How big is your lawn?”, I would get a variety of responses.  Many common answers are “1/4 acre, 1/2 acre, 10000 square feet, or maybe 900 square yards”.  The answer I would not expect to get is “45 minutes”.

Why is that?  For one thing, I asked for a size estimate, not a time estimate.  Yet, all over the globe, if I ask a software developer or project manager to estimate the size of a feature, they will generally answer in a number of hours, days, weeks, months, etc.

Time instead of Size.

This behavior has been ingrained in the development community for decades.  Agile methods have attempted to break that habit by requesting effort estimates as a measurement of relative size (often story points).

Indeed, I have been coaching and guiding teams long enough to have seen a variety of attempts to shift teams to “how big?” rather than “how long?”. Admittedly, it’s a really tough transition to make. And a very common way of approaching that transition is to offer a crutch – usually a conversion factor … you know, so the team members can wrap their heads around the concept more easily.  A common one I’ve seen is 1 Story Point = 1 Ideal Day.  SAFe offers the crutch that 1 Story Point is about 8 hours.  I’ve seen someone recently amend that to a range from 1 to 10 hours.

I’ve adopted a lot of truisms in my years of observation and continuous improvement. One that has stuck with me consistently though my agile career, is simply this:  “There is no direct relationship between Story Points and Time.”  If I want to leave some wiggle room, I might say “no reliable relationship” or “no consistent link”.  Of course then I’ll toss in that story points and time is not constant from team to team.  Any attempt to build that bridge will be fraught with drama, and lead to disappointment.

But nobody believes that.

And so, again and again, those who won’t acknowledge the past, repeat it over and over.

I’m hoping this series of articles on Agile Estimation and Planning will serve to save even a handful of you from following so many others in walking off that cliff.  If nothing else, it’s a chance for me to hone my explanation.

Feedback is welcome, please weigh in.

Part 1 – Size

Let’s go back to the lawn example, and start with a simple explanation as to why time is a terrible way to measure size…

I asked you how big your lawn is. I want you to think of your lawn. Think of the dimensions, think of how much of it is actually available for mowing.  Yes, that’s right. The grass.  Not the driveway.  Not the garden, nor the planting beds.  Not the patio or the stoop.  Got it?  First of all, notice that the size of your lawn is not the size of your lot.  The lot represents the dimensions of your property.  But not all of that is lawn.

Are we good so far?  I don’t want to lose you.

Now.  Picturing your lawn, and only your lawn:  How big is it?  Odds are, unless you’ve played this game before, you can’t give me an exact answer. You probably could if I gave you enough time and a tape measure.  You could probably get me a pretty exact measurement of the lawn part of your property.  But I have two problems with that.  First, it would take too long, and second, that wouldn’t be an estimate anymore, would it?  It would be an actual.  Traditional estimation models excel at doing this (pun intended), by giving you a really long time to lull yourself into a false sense of security in your ability to predict the future.  You may have noticed this activity is often followed later by a check of how accurate your estimate was — or more to the point, how wrong you were.  “We gave you months to predict the future! Why were you unable to give us a better estimate?”  That’s a hell of a way to start a conversation, isn’t it.  Let’s try to accentuate the positive a little, shall we?

In a moment we’ll get back to talking about the size of your lawn…by comparing it to something else!  Studies have shown that while humans are pretty bad at exact estimates, they are very good at comparing things.  This is a survival skill you learned as a small child – especially if you had siblings.  Remember family deserts?  Remember when there was one piece of cake left and both you and your pesky brother (or sister) wanted it? Remember how your mother, with the wisdom of Solomon, suggested you split the piece? The rules were simple: One of you sliced.  The other one chose – which gave you all sorts of incentive to slice well.  But no matter how carefully the slicing was carried out, one piece wound up slightly bigger than the other one and I’ll wager my half of the cake that you could tell exactly which half was bigger (the one you wanted), just by looking at it.  Or perhaps, they were actually close enough that there was no real distinction between them, and you both got to eat a piece of cake with satisfaction of knowing nobody got cheated.  We are going to leverage this skill in Agile Estimation.  The ability to look at two things, and declare whether one is bigger or smaller will prove valuable here!

Ready?  Do you need some cake first?  We don’t have time for cake.  Cake is for closers.  Are you a closer?  Let’s see what you’ve got.

Think about your neighbor’s lawn.  Let’s see if you can answer a simple relative sizing question.  Is your neighbor’s lawn, bigger, smaller, or about the same size as yours?  If you live in the suburbs this part will be really easy. Odds are pretty good that your neighbor not only has the same size lot as you, they probably also have a similar amount of grass in their yard.

So, what’s your answer? Is one lawn bigger than the other? Or perhaps, they’re close enough that you’re willing to call them even.  We’re talking now about degree of precision. When we ask for estimates, we should not be asking for perfection.  We just want to know within a margin of error.

To make this next part easier, I’m going to switch to talking about myself for a sec, partly because I’m having trouble seeing your lawn, but also because my neighborhood provides a particularly compelling example.

I can say with a fairly good degree of certainty that not only do I and my two adjacent neighbors have similar-sized lots, but the amount of open grass is pretty similar between the two — close enough at any rate.  That means, in terms of relative sizing, our three lawns are the same size.  How big is that?  Not sure yet, let’s think a bit farther.

Now think about your neighborhood, and the finite universe of houses and lawns that surround you.  Can you think of any lawns smaller than yours?  Can you think of any that are bigger?  If you can answer ‘yes’ to either of those questions, then the next question will be “How Much Bigger, or How Much Smaller”, and because we haven’t established a unit of measurement yet, let’s just use relative sizes.

Returning to my yard, the houses in the middle of the block all have the same lot size.  I also can’t think of any houses in the area with a smaller yard; but I can think of a few yards on the corners that are bigger… almost twice as big!  And then if I think about all the houses on the block, there is one house that due to the curve of the road, I’m pretty sure his lawn is closer to triple the size of mine.

I’m going to establish an arbitrary starting point now.  Based on my understanding so far, my yard and my neighbors’ are all 1 point yards, the corner houses have 2-point yards, and the odd-shaped yard a 3-point yard.  The chief advantage at this point, is this measurement was FAST.  Not a lot of time spent agonizing over it.  It was a quick 1,2, 3 and we’re ready to move on.

Figure 1.1

I now have three data points we can use to compare other work to.  If someone asks me over to their house and say, “could you estimate my lawn?”, I need only look at it, and compare it against the three data points in my head.  Is the new lawn bigger, smaller or about the same as one of those other lawns.  And if it is bigger or smaller, how much so?

The other advantage is that this estimate is PORTABLE – meaning it can be applied to the work of more than one team.  We’ll start to see why that’s important in the next section.

Part 2 – Scope

At this moment, all I have is a cursory understanding of how a few pieces of work relate to one another.  There are a myriad of activities that one could perform in support of a lawn.  We’re going to focus initially on mowing.

One weekend not so long ago, I looked at a long list of activities I needed to work on, and was trying to figure out how much I could do with the time I had.  One of the things that I have a pretty good handle on is mowing my lawn.  My definition of “Done” was pretty basic: gas up the mower, make sure it was set to “mulch”, and wander the property until all the grass was a nice, mostly uniform height. On most weekends, using my self-propelled lawnmower, I have been able to complete that job in about 45 minutes.

“Ah HA!”, I hear you cry, “Time!  Your one point lawn takes 45 minutes.  I’ve cracked the code!”

Well.  That’s nice.  Good for you!  But before you pat yourself on the back and start dividing my backlog into 45-minute increments, I’m going to wrinkle this up for you a bit.

Last summer, I hired someone to mow my lawn for me.  Don’t judge.  Work was demanding, and it was hot, and once you hire one of those guys, they keep coming back.  So my point is this:  Ed didn’t have a measly self-propelled walk-behind mower.  No.  He had one of those zero-turn-radius, stand/ride-behind monster machines that goes zero to sixty in about 2 seconds.  He fired that bad boy up, and viola! My lawn was freshly shorn in 10 minutes!

So, let me ask you this: Did my lawn change size?  If there is a universal constant between time and size, then Ed’s performance had reduced my 1 point story to something under a quarter point.

I submit to you that my lawn is in fact, the exact same postage stamp it has always been.  If it was 1 point before, it is 1 point now.  But something is clearly different.  Even I have to admit that 45 minutes vs. 10 is pretty compelling (or at least disheartening).

The difference has very little to do with the lawn itself.  The difference is that Ed, with his superior machine is capable of delivering the same work in less time than me.  How can I use this?  Well, if you take my 1-point lawn, and assume that every approximately 1-point lawn will take the same approximate amount of time (for me) then I can use that understanding to suggest how many 1-point lawns I believe I could fit into a weekend.  Given what I know, I think I could complete approximately 16 points worth of lawn .  But my friend with the impressive hardware can do many more!  The math whizzes out there have probably already concluded that Ed can do six 1-point lawns in an hour.  So given an eight hour workday and two days in a weekend, he’s going to be pounding out 96 points worth of lawns every weekend.  Those same math whizzes will point out that my sixteen-lawn estimate for myself is obviously under-represented.  I should be able to do four lawns every 3 hours, so in an 8 hour day, I should be able to do 21 points worth of lawns in a weekend, and if I just did a little overtime, I could pull off 22!

Isn’t math awesome?

and Terrible?

Allow me to toss a little water on those flames of victory you’re dancing around.  Unless you are going to work things out so those 21 lawns are placed end-to-end next to each other and set up my mower to be perpetually full of gasoline, I can’t possibly attain the number you’ve so carefully assigned to me.  I have to move the mower from one location to another.  I have to refuel at every job.  I have to stop to take a drink of water, or eat lunch, or answer a call.  Your assumption of my velocity seems to be missing a few things.  Your assumption of Ed’s velocity is just as wrong, by the way.  In order to move that behemoth of his, he needs to drive it up onto a trailer in order to move it to the next job site.  Where would we account for loading and unloading time?  Where do we lump the travel time between locations?  What about bathroom breaks?

The capability of a team is made up of a lot more than just the sum of the hours of work they perform.

Part 3 – Definition of Done

Something else happened when I hired Ed to mow my lawn.  He brought friends.  Not only was Ed mowing my lawn, but he had another guy with a gas-powered trimmer and yet another guy with a leaf-blower helping him.   So when he was “Done” with my lawn, not only had he completed the mowing job in less time, but he had DONE MORE than I ever did in my 45 minutes.  Trimming!  Edging!  Cleanup!  And once my wife realized those were on the table, those all became desirable additions to my lawn mowing regimen as well!  It was no longer sufficient for me to just gas up my mulching mower and try to get as close as I could to the flower beds.  I needed to pull out my electric trimmer, and follow myself around the yard, then edge the sidewalk, and then make sure I swept the sidewalk and driveway.  Guess what!  My 45-minute job wasn’t 45 minutes anymore!  Ed has a team of three people.  My team still only has me on it.  All that additional work added another hour to the time it takes me to complete a typical 1-point lawn (thanks a lot, Ed).

I ask again.  Think carefully.  Has the lawn change size, now?

No, the lawn is still the same lawn it always was.  But the definition of what constitutes a completed, quality job has certainly changed!  And because I now need to do more things to complete the same sized lawn, rate of delivery will suffer for it by decreasing.

Understanding this distinction is absolutely key!  The job didn’t change size, the level of expectation changed, and my team failed to adapt to that change.  The result was a measurable impact on my ability to deliver value (completed lawns).

Why would I insist on looking at it this way?  Consider this:  Recall, earlier I had estimated that I could deliver 16 points of lawns in a weekend — sixteen lawns just like mine.  In order to compete with Ed’s lawn service, I can no longer get away with a mow-only task list.  I need to perform the same quality tasks that Ed provides!  If I equated size to time, I would now have to re-estimate all of the lawns in my list!  Instead of 45 minutes, they’re taking 45+60=105 minutes.  By that magic conversion factor of 1 point = 45 minutes, those 1 point lawns in my neighborhood are now 2.33 points each (because math).  The corner lots are now 4.66 and the weird lot is a whopping 6.99.

Also, since you already had the calculator out, the project manager/mathematicians went even further: “By my calculations, in the same 16 hour work weekend, you can now to do (16 hours x 60 minutes per hour ) / 105 (minutes per lawn) = 9 lawns. 

See how easy this is?  You changed the scale and did math again, trying to get me to commit to something like (2.33 x 9 = 20.97 — aw heck, let’s say 21 points).  When a 1-point lawn was 45 minutes, you demanded 21 points.  Now that a 1-point lawn is 105 minutes, you still want 21 points.  Except that before, that 21 points would have delivered 21 pieces of customer value, and now that same 21 points is only delivering 9 pieces.  My team appears to have maintained velocity (that’s good, right?), but we’re not advancing anywhere near that rate!  Only 9 customers are satisfied by the same 21 points.

Also keep in mind, that I’m still not accepting your math.  I was only willing to commit to 16 points per weekend.  And I’m not going to re-estimate everything on my backlog, either.  THE LAWNS DIDN’T CHANGE SIZE!

With the impact the new Definition of Done is bringing to the table, unless I do something to correct my team makeup, you’re going to see my velocity drop like a stone to 7 points – seven lawns instead of sixteen.  And that very rightly should set off a red flag somewhere.

Part 4 – Velocity

When we talk of delivering value to our customers, we need to be very careful that we only count things that we’ve actually completed. The Definition of Done says I must mow, trim, edge and clean to meet my customer’s expectations. If I fail to deliver on any single aspect of that definition, the customer could reject my work, and refuse to pay me.

Think about that.  Even if I spent every minute of my weekend pushing the lawnmower around, and managed to cut the grass on twenty lawns — if I never touched the trimmer or broom, then despite the fact that I was BUSY the whole time, I didn’t actually COMPLETE any of the work.  What do I do now?  What do my customers do?  Do I just come back next weekend and do the trimming, edging and sweeping then?  Nice theory, but the grass will all have grown back by then, so now I’ll spend all weekend finishing he jobs from last weekend, but the customer will look out and still see incomplete work, and they’d be well within their rights to withhold payment again.

This is why the Agile world is OUTCOME-DRIVEN. It doesn’t matter that we are keeping our people active and working (busy) if they aren’t finishing everything in the Definition of Done.  When you get right down to it, time estimates are not helpful.  We don’t deliver time.  Our customers don’t consume hours.  They really don’t care how busy we were.  They care about results.  So we need a measurement of our ability to deliver VALUE.  In Agile, we call this Velocity.

Velocity is a measurement of the number of points of value that a team is able to complete in an iteration.  Only work that meets the Definition of Done is counted toward Velocity.  Partially completed work may as well not exist.

Velocity is a very useful metric, but it is not a universal constant.  It is a value that is unique to every team, based on their skills, tools, team structure, etc.

No matter what, Ed’s team is going to have a higher velocity than me.  First, their equipment is better -I can’t compete with that monster mower. And second, his cross-functional team can do tasks simultaneously.  In terms of clock time, from the moment they pull the equipment off the trailer until the time they roll it back up, something like 15 minutes goes by… (this includes Ed coming to the door for his payment).

Despite the fact that I’m telling you not to estimate in time, the concept of time keeps working its way into the conversation.  For instance, if I say Ed’s takes 15 minutes from wheels down, to wheels back up on the trailer.  That’s time, isn’t it?  You’re right, I did talk about duration of the job.  But I didn’t ESTIMATE in time.  Do you understand why?

The reason we want to stay away from time estimation is because time estimation is not portable!  It cannot translate from one team to another. When Ed takes 15 minutes, or I take 105 minutes, this causes a problem when it comes to translating our work.

My hours are not your hours.  The way I work, and the experiences I’ve had will ultimately affect the amount of time it takes for me to deliver.  What if Jeremy, the kid across the street grabs his dad’s lawn-mower and decides to get into this lawn-cutting game for some extra pocket money.   As it turns out, Jeremy’s mower is exactly the same as mine because his father and I both bought them when the local big-box store had a killer overstock sale.  Jeremy and I have the same equipment at our disposal.  Does that mean he is going to take the same amount of time that I will to mow my lawn?  I will bet you the answer is NO.  Whether Jeremy and I have different levels of strength, speed, attention span, or even experience, I can virtually guarantee that we will perform the same amount of work at different speeds.

Furthermore, if Jeremy or I walk up to a house on the next block, intent on giving our estimate, and instead are informed that Ed was already there, and told them it’s a 15-minute job.  Does that mean I’m required to accept that estimate?  How many times in your professional career have you found yourself being held to an estimate made by someone else?  It happens all the time – even more disturbing when that estimate isn’t even made by someone who does that work.

Part 5 – Cross-Functional Teams

Let’s assume you’re someone who’s going to fund the mowing of lawns.  Maybe you’re the head of the homeowners association, and you’re in the process of subcontracting the upkeep of lawns in the subdivision.  My team originally pledged 16 points, and my competition pledged 6o points.  But after you saw the additional services they were providing, you requested that we have a uniform Definition of Done across both teams.  This threatens to tank my performance!

How could we reconcile this?  “What,” you may ask, “do I need to do, to maintain the higher committed velocity?”

I suppose I could throw caution to the wind, and try running behind the mower!  I might be able to bring the time it takes to mow down a bit, but I’m probably going to miss some spots.  I don’t think that will ultimately solve the problem.

I could work overtime, mowing long into the evening hours.  That has potential, but now I’m going to get way more tired, and likely make mistakes.  The quality of my work will certainly suffer, and I now incur the risk of the customer rejecting the work. What would you do?

How about adding people to my team to handle these other tasks in the Definition of Done.  Ed has a cross-functional team (mower, trimmer, sweeper), where I had only a single generalist on my team).  If I got someone to run the weed whacker while I mowed, and whichever one of us finished their job first, then swept the sidewalk, I might be able to bring my clock time down to under an hour.  The math would support us saying we could reach 8 points per day then, but I’d still feel a lot better calling it 6 or maybe 7.  Either way my Velocity for the weekend could increase from a around 7 if I go it alone, to a real possibility of 12-14.  Not quite 16, but let’s face it: when I was just mowing the lawn, there was still long grass growing against the fence and the planting beds that the mower just couldn’t reach.  There was still grass clippings on the sidewalk, and let’s not even talk about the edging! In short, my quality was low, and the other team revealed that fact to the stakeholder.  I had to adapt or risk losing the gig.  So I expanded my team.

This notion of a cross-functional team is very powerful.  It allows us to build teams with all the skills needed to achieve our Definition of Done in one iteration, without relying on outside help.

Part 6 – Multi-Tasking and Work in Process

How good are you at multi-tasking?

Part 7 – Stable Funding

…or more importantly, our value.  What if I want to find out when I can expect delivery on the 16 points of lawns Figure 1.1?  It’s a valid question.  It happens all the time in business.

 

 

Are We There Yet?

A friend of mine was once put in charge of the Innovation group at the company where we worked.  Whenever we would talk about the state of Innovation, he always seemed preoccupied with creating an “innovation organization”.  But the conversation never seemed to get far beyond that point.
Continue reading “Are We There Yet?”

Accentuate the Negative?

The instant message app chimed on my desktop. “How confident are you that we won’t find any more defects in testing?”, the department head asked. I glanced over at the task board, at the lone sticky note sitting in the “In Progress” section: “Architecture Review”, it read. I popped open the chat window, and responded, “100% sure. We are done with testing.” I watched the window for a moment as the message app informed me they were typing their reply. “Okay. Because I’m about to go report that to the senior managers.” “No problem,” I typed back. “See you at the demo, tomorrow.” Continue reading “Accentuate the Negative?”

Give a Boy a Hammer…

“If you give a boy a hammer, he will suddenly find that everything looks like a nail.”

In some ways, this is how Agile is being applied in the industry today. It doesn’t matter what problem you need to solve, hit that nail with a Hammer. You want to get better predictability? Hammer. Quicker time to market? Hammer. You want to improve employee morale? Hammer. Continue reading “Give a Boy a Hammer…”

Agile vs. Waterfall – Improved Performance is NOT Guaranteed

I am frequently asked to give a brief overview of Scrum to people who are unfamiliar with Agile concepts. In the course of giving those lessons, I almost always see a look of shock at the almost cavalier way that we agilists claim that Agile methods will give a better result than traditional methods. I like the look of shock. It shows that they’re paying attention. Continue reading “Agile vs. Waterfall – Improved Performance is NOT Guaranteed”

On Tweaking Estimates

As Product Owner for the project, I get to field a lot of interesting questions. I woke up this morning to the following question from one of my teams:

Hello Mike,

In our current sprint, we have planned 8.5 user story points. But after task breakdown we can see for a couple of user stories, actual story points differ from the estimated story points. What is the correct course of action in this case? Do we need to update product backlog for revised user story points? Then we can say the team is working towards achieving 11.5 user story points! Also, since our [project gate commitment] is nearing, should we reexamine the estimates in the product backlog, and revise them for the remaining user stories?

Continue reading “On Tweaking Estimates”

Sailing Metaphor for Agile

When I was in high school, my friend Scott and I took his father’s sailboat out on the lake where our families spent summer vacation.  We had each been sailing numerous times; always as crew, never at the helm.  We both knew the lingo; “Come about” and “Pull that jib in tighter!”.  I had taken a boat safety course, so I knew all about life jackets and right-of-way.  There was even a page in the safety manual that talked about sailboats and gave a handy chart for the points of sail.  We were set for adventure! Continue reading “Sailing Metaphor for Agile”

The Culture of Agile Adoption

As a self-proclaimed Agile Evangelist (Agilist), I am keenly interested in the process of taking an organization from a traditional project management base, into a truly agile model.  There are a variety of factors that come into play that make the transition from traditional to agile more of a challenge than it first appears. Continue reading “The Culture of Agile Adoption”