Flash “Mobs” & #NoEstimates in Manchester
Walking into the Lion’s den
by @EtharUK
Rantosaurus
If anyone followed my previous blog posts and twitter rants over the last few years, they’ll have noticed the evolution of my very hostile stance towards the #NoEstimates movement. Whilst I take a critical look at everything, I like to think I am definitely not closed minded. It’s very rare that there is nothing of any value [to someone] in something, so I always take an exploratory stance on everything, and categorise by best fit.
Now, whilst folk like Woody Zuill and Nick Killick have got on my proverbial wick more than once, especially after the first couple of interactions finally had me insulted :-D I want to reiterate that the method isn’t without it’s merits, but it has to look at itself in it’s limitations.
- It does not work for high process environments, you are not given the space for it.
- It requires an absolute dependency on a mature, agile delivery skill.
- It requires folk to understand when in the cycle it’s most valuable (which is often the place most managers are most scared, so ask for estimates) and
- When and when not to play to that
To be clear, being the sort of guy I am, I have absolutely lambasted Woody in particular. I mean REALLY lambasted him! I got insulted the very first conversation we had and it went downhill from there.
At one point, I even stopped referring to him with this first name when discussing #NoEstimates with other folk. He became “THAT F**KING ZUILL!” in conversations.
Disenfranchisement
Despite embedding myself in the #NoEstimates movement in intense periods for the best part of 2.5 to 3 years, I eventually got fed up with the distinct lack of progress and the weaknesses in the positions put forward by #NoEstimates advocates.
The hashtag was, by far, the biggest barrier to any forward progress, because for each step #NoEstimates took, they had to stop and try to convince 1,000 people, using really poor technique, a lack of shoe-walking and approaches to communication, that #NoEstimates doesn’t mean “No Estimates”. Not least because at the same time, it failed to limit the idea of an estimate to time or other quantitative measure as it’s used by people.
Remember that point, as I will refer back to it. That is a massively important constraint on the definition, because every hypothesis is an estimate! Every business case for a product or service, is an estimate. Many, if not most, people’s roles exist precisely because of a hypothetical business case. Indeed, even the direction that the #NoEstimates movement itself advocates is based on a hypothetical estimate. “We do what we think we should” can be translated as “We ‘estimate’ that doing this thing is better”.
Business Case: “Carry on making our current software work better, so we don’t end up losing the farm. We need developers for that”
Business Case: “If we do this thing, this will bring our customers the benefit of… We need developers for that”
To [un]ashamedly one of Woody’s statements against him “Do the most import thing until it’s not the most important thing”. This estimate of “most important” is subjective. Where organisations are not working lean, the systemic flows mean that you often don’t have sight of the most important thing, only your small part in that chain. Your most important thing isn’t aligned with what’s most important to the enterprise’s end goal. So your most important thing is, and can only ever be, an estimate of the most important thing. It requires someone else to show you that it’s not the most important thing.
Every single UX, A/B/n-test or refactor that we do is based on an estimated benefit. So the definition has to be precise which is a theme that runs through the #NoEstimates debate time and time again. It works in a very specific set of circumstances and that’s cool. It’s a tool. it goes into a toolbox. Like a mis-set bone, two step refactor, or more widely, the last few steps solving a Rubik’s cube, you have to break it to make it better. However, the personalities did not lend themselves to that way of thinking at all. Indeed, many NE proponents absolutely lambasted measurements in general, throwing the baby out with the bathwater, akin to how poor agilists did in the very early days. So the conversations got fractious, very fractious.
To me it indicated a very very poor understanding of Russel Ackoff’s famous quote:
“Managers who don’t know how to measure what they want settle for wanting what they can measure.” — Russell Ackoff
I’m going to ask a question and leave it there. Is that a problem with measurements or managers? If you think it’s measurements, why are you producing the measurement? Is your boss telling you they need it? Are you the measuring tape? What value do you get out of it?
Given I’m a huge proponent of data driven… well, everything… I’m big on the idea of measurement. However, as someone who’s that heavily into it, I’m also of the opinion that you should understand the question you want answering before finding the method from your repertoire of tools to measure it. It’s perfectly satisfactory to be “just good enough”. For example, if you are adding numbers of 20 decimal places of precision for calculations only to round them to pounds later, then it will rarely affect your accuracy if you moved to 19 decimal places or even 5 if there aren’t that many numbers. Knowing what to measure is a balancing act between having a wide set of tools and techniques to measure, your ability to discern which tool to use (with their limitations) and measuring just the right amount of the right thing.
Why I Went Along
At this point, you’d be forgiven for asking why the heck I would put myself through that? Given I hate the guy and what the movement had become, I’d be foolish or given I grew up on the wrong side of the tracks, I went to start a fight.
The event was a flash meetup. There was only 2 days notice and Gemma Cameron, Manchester’s resident tech event organiser, ferret owner, chees joker and coder herself, did a sterling job organising attendees and finding a venue in new The Cooperative Group building. at One Angel square.
The reason I went along was several-fold, but crucially, I wanted to understand less than if Woody’s position stemmed from complete and utter ignorance and stupidity, and more whether there was common ground. It was meant to be a positive experience and I hopefully made that clear on twitter a few days before.
I wanted to know if it was one of those situations where, at a fundamental level, we were close yet approached our journeys differently from then on. Plus, I have enough self awareness that I can violently agree with people, so wanted to check, in a forum that wasn’t twitter (because it’s rubbish) and without the confounding influence of others, that we had some common understandings. We came from the same desires and passions. I was even prepared to apologise if I had to.
Conclusions
I managed to keep quiet. I did it! WOW! Asked one question only. First time in my adult life! Plus, I finally drove away at about 10:30pm and got a kebab on the way home…
Oh you want to know about #NoEstimates? OK. Woody’s baseline position is fine. He and we do come from a common set of aims. He stumbled upon both Mob and NE when working with teams who were already using good practises in their work. Indeed, his mob programming talk sparked something.
“Nobody should be in a team if they are not contributing their knowledge.” — Woody Zuill
Turning this a little on it’s head, this leads to a possible solution to the plateau effect often seen when mob programming is done. The teams start of really well, but as the knowledge transfers through the team and/or the work becomes more trivial, less and less individual contribution is made by each person, as that knowledge is then known and shared by more people. This means those folk are paid to just sit in a team not contributing.
Using that approach allows you to understand when people can be freed to move to other teams. If the same value is being delivered with less people, it’s a conversation piece with a “resource manager”. Any superfluous individuals are free to find another project to contribute to. However, this is a culture change on behalf of both parties. The organisation and the individual especially to say when they feel they’re not contributing.
Also, I am heartened to note the crucial “admission” by Woody. Estimates aren’t the problem, it’s people. People hold other people to estimates which are very often rubbish. The issue isn’t the estimates there, it’s the people. Getting better at estimates also isn’t the nub of the argument for folk like me, it’s the delivery of value.
How fast can we get something working and into people’s hands so we can learn from it?
Just like any project or programme management probability, which is mindful of probability (T-shirt sizes) and impact (T-shirt sizes), there are two dimensions to it.
- Speed
- Something working
- Learn about the customer’s use of it
This is why thin story slicing is so valuable. The cone of uncertainty is widest at the beginning. It tells you the chance of you hitting that value is slim. The thin stories, deliver something working, quickly that you can learn from, and give you more certainty. Where have you seen this breakdown before?
Personally, I’m still not convinced folk in the #NoEstimates movement as a whole have the right approach. This still has a lot of work to do to convince a lot of people and to engage them in a way that works. At the moment, it doesn’t engage people. Indeed, it makes them downright hostile. However, it was good to see that Woody does come from the same baseline as ourselves and shares the same goals. Plus, at least he is conceding points he wasn’t clear or right about previously. I’m good with that. So I can only offer him the same courtesy.
Mr Woody Zuill, I’m happy to use your first name again now. We did the formal apology yesterday.
Liked this article? Don’t forget to hit the heart. Also, follow BzSkits for more real life digital business.
