In the first part of the series on agile retrospectives, I mostly focused on the visual inventory as I went over the reports and charts we've been using at TargetProcess, along with some stuff from other people. I've also given an outline of the heuristic, trial-and-error approach as the essential methodological foundation for running retrospectives.
In the second part, it’s time to look into the secret nuts and bolts of what actually makes retrospective meetings work. I’m stepping out to embrace a broader picture, as the subject of company culture — it’s exactly about the quality of this oil that makes the retrospective engines run — is limitless. It can’t be reduced to a few worn-out, cliched how-to adages.
Maturity and the culture of trust
A team has to be mature enough to hold retrospectives. One of the common pitfalls is when teams are eager to run retrospectives just for the sake of running retrospectives. They know that if they are supposed to be agile, they need to do retrospectives. But it’s not as simple as following a guideline. The need has to come from within. The team has to develop this intrinsic feeling of power to solve their problems and - even more important - to accept the responsibility.
So how would you know if a team is mature enough to run meaningful retrospectives? If it’s not, how do you go about nurturing such a team? Do you have to bother at all?
First and most, it’s about the no-blame culture. Well, it’s better in the positive, not in the negative, so let's put it like that: no-blame culture is the culture of trust. Let me mention here that trust is one of the signature values in our company.
The culture of trust emerges from a dozen of subtle ingredients. There’s no exact recipe as to how much of each ingredient you should take and in which proportion. It’s not a cocktail, for better or for worse, but it does bear some resemblance to a cocktail, as oftentimes you've got to stir, or even shake it well before enjoying!
Communication: tell me who you are, and I’ll tell you who I am
Following the lines of cocktail making, communication appears to be the basic ingredient of the culture of trust. I completely agree with the point of this article that “lack of communication” is an umbrella term for all kinds of hidden conflicts and unresolved issues. It represents an antithesis of the culture of trust.
So, if in your company “lack of communication” is light-heartedly referred to as the common excuse for troublesome issues, that’s when the red lights should go flashing, better yet with the fire alarm sound. When people don’t talk, or misunderstand each other’s messages, or are reluctant to spend time explaining their point of view– this is the surface skim of what can be a severe turmoil of contradicting values, assumptions and beliefs.
For a certain time, this turmoil might simmer as a dormant volcano, but the eruption can come at a bad time, when everyone assumes the volcano is asleep for good. At a retrospective meeting, for example, in the form of finger pointing:
Finger-pointing, it seems, falls short among the other consequences of the notorious lack of communication. Look, if tons of paper and web pages have been wasted to research/discuss this problem through and through, but it still persists — it means that avid readers are missing the point in their own organizations, being either short-sighted or far-sighted, or sometimes even both. It also means that some essential things are left off from those writings, as for some reason people fail to fix the glitches once and for all. That’s because they don’t get to the bottom line.
For now, I’ll take just one cultural pitfall related to retrospectives and meticulously dissect it, tissue by tissue (surgical metaphors seem to fit better here than cocktail recipes).
The pitfall: no one wants to attend retrospectives
All right, about this one. Some common tipsy-tricksy symptomatic cures and surface reasons for this pitfall I’ve seen in some books include:
- people don’t want to spend too much time on a retrospective, that’s why they rather stay away, so let’s set a time limit of 1 hr (insert 0.5 hr etc.), and this will fix it
- no one wants to attend because it’s soo boring, so let’s make retrospectives fun! Let’s go somewhere outside and invite some clowns so they entertain us, or let’s have gourmet chefs cook delicious food for us. Only then a retrospective will work.
- people don’t feel relaxed enough. They’re shy to speak up, they experience social awkwardness, so let’s do a group therapy session with a set of relaxation exercises.
How does that sound? Shallow, at least. So, I’ll run the chain of whys to find the underlying deep reasons (see the first part for another example of the whys technique).
No one wants to attend retrospectives. Why?
They believe this is a waste of their time. Why?
They decided to take some actions at the previous retrospective, but it didn't work. Why?
Everyone seems to be fine with the current setup. Why?
... and so on. Most typically, the chain of whys will uncover one and the same issue: a retrospective is not a meaningful activity. Sometimes, a retrospective is just a ritual the team leader or the CEO is following to do agile by the book, while in reality they're well aware of all the problems, and already have some solution in mind. So, a retrospective in this case is more of an opinion-check meeting to sense the atmosphere in the team, and to see if the leader’s decision is aligned with the expectations of the team or not. Of course, it’s not all black and white. The authoritarian leadership style is hardly viable anymore, but some annoying breadcrumbs of this style can cause cultural itching. Nor an utterly team-based ruling will make it work ultimately.
A report from the trenches
I've already mentioned briefly that we're now working on the new version of our product, and we're back to doing retrospectives. Recently we've had an intense 16-day span called Subbotnik to come up with what turned up to be a private pre-beta (not even a beta so far).
So, there was a retrospective, and the team along with the Scrum Master decided that we're now doing it in a milder mode, with good old 2-week iterations, and even with a burndown chart. Also, there appeared to be a deadline (something unheard of in our company for several years), so people somehow decided that we need to suspend the IDP (Individual Development Program) until we release for the deadline.
A side note: we have an individual development program, similar to Google's famous 20% time, and this is one of our signature practices based on the essential company values. Maybe the Scrum Master was too rigorous in propelling his methods to reach the desired outcome, or the team believed that this deadline is indeed a matter of life and death, either way, they decided they’d suspend the IDP until that deadline. There were signs of the latent simmering, and after some get-to-the-bottom-of-it discussions it’s been finally made clear that we are NOT actually chasing any particular deadline, we just need to do it as fast as we can, until it comes out good (in short), and we will not sacrifice the IDP practice to this deadline.
This is a very characteristic example, a perfect specimen of “the lack of communication”. If the team and the Scrum Master were aware that there’s no deadline actually, they wouldn't have opted for holding off on the IDP, as much as it means to our company. So, the lack of clear communication caused a slight shake to the cultural foundations, which later on might have developed into the subdued discontentment.
The core of this deadline/no deadline problem lies in something personal. It’s about the desire to come up FASTER with a great new version of our software, the software that is so incredibly helpful indeed. If speed becomes an obsession, it can backfire with the unpredictable side-effects. Anyhow, in our case, the urge to deliver fast collided with the higher values, besides, we've learned that deadlines might only work for Subbotnik (the development practice we’re using to clean up residual bugs).
More analysis: we’re a company with the culture of trust. So, the issue of finger pointing hasn't been affecting us much, from what I see. But at times some communication glitches occur. Even with trust established as the core value, lack of communication might produce the effect of hidden blame placement. Generally speaking, for any production work, the people that one wants to criticize most are those who do not deliver enough information in the context of tasks done together. This non-delivery of information, I think, is not an intentional evil act, it happens inadvertently. If someone is obsessed with speed, it wouldn't occur to them that by keeping the latest update or idea to themselves they’re doing anything wrong. This is the case of too narrow focus, or maybe being too far-sighted, letting people out of the mutual loop.
You might ask: so what, what’s the solution? Is it just one more story about the lack of communication? Is there any real cure to it, other than the recipes fit for kids daycare? Note, that we’re still talking about agile retrospectives based on our IDP non-suspension case. The real cure is to get the point through to people who set the outer culture and the inner culture. Lack of communication would continue to be referred to as an all-forgiving excuse, until someone who is in charge truly gets that the side-effects violate company values AND in the long run slow things down, not speed them up. That’s a paradox, very well known from the story of the turtle and the hare: Festina lente OR “Hurry up by being slow”, as translated from Latin. Things will not happen faster if you save those few seconds on not mentioning an important update to people who work all together, especially in a location that high. This famous picture brightens up the doorway in our office, along with some others, and I find it particularly fascinating.
To put it in other words, no cultural attitudes are shifted by a compulsory decision to follow some prescribed “shoulds”. The attitude might only change if an established behavior pattern is violating some higher meaningful values. I’m not giving any “how to make bad patterns change" advice intentionally, nor saying what those higher values could be. There’s no universal answer and recipe. Every organization has their own secret nuts and bolts, and in the end it all gets down to people, their personalities and interactions. Forgive me, my dear techies, but it’s about psyche and emotions, not about stats, graphs and calculations. Binary thinking with its universal “wrong/right” -“good/bad” rhetoric has no answer. There’re countless subtleties and nuances. The point is to be able to identify the prevailing idea or sentiment or value, and go from there.
Here’s a simple life-based example, not related to software development or project management, just to show how “need” or “can’t stand this anymore” beats “should”. For my tennis practice, I have a set of balls, usually about 60 or so. They last for about half a year, sometimes longer, and then they need to be replaced with a new set. My latest set of tennis balls has been alive for almost incredible 12 months, and it suited me OK, because I’m practicing on the fast surface, and while old balls don’t bounce as fast as new balls (well, there’re different varieties, but this particular set had exactly that aging pattern), it’s almost like the imitation of the slow clay where you have more time to get to the ball and to make a good move. The fast bounce has some disadvantages, to cut it short, and it's not what I'm focusing on at the moment.
So, my coach kept reminding me that the balls should be replaced, that they're old and they bounce slowly. I kept dragging, finding excuses not to replace the balls, although, of course, I should have had them replaced weeks ago. Guess what did the trick to me? The sound. Old tennis balls produce this particular “rotten” sound, and with my sensitive musical hearing, there came a time when I finally couldn't stand it anymore, and took care of the replacement without further lingering. So, you can see that the “should” didn't work, not until a higher value was challenged.
Single-Digit Teams and Mainstream Advice on Retrospectives
The closest I can get to a “how-to” style is as far as saying that a single-digit group (<10 people) is less likely to have communication problems, and more likely to act sensibly as a team, at retrospectives as well. But again, organizations change and grow, and what has started as a small start-up becomes a mid-sized well-established business.
Back to the tips and tricks I’ve mentioned above. I can’t say that they’re good for nothing at all. Quite often, in a larger organization, there’s no real lever to influence culture shifts for people assigned to run retrospectives. If someone is confined in a certain space, like, a singled out task of doing a retrospective – then they would look for the ways to create some ado around this ritual, and operate in the limits for which they’re accountable. Role play, shuffling moderators, allocating fixed times, inviting clowns etc. — such advice is tailored to this very audience.
But I hope the awareness is there. The rituals do not set in motion the really powerful driving forces in a company. That’s what I’ve tried to outline in this part.
Next in the series: Joy Spring and Estimated Deadlines