This is a transcript of episode 211 of the Troubleshooting Agile podcast with Jeffrey Fredrick and Douglas Squirrel.
Squirrel and Jeffrey discuss how to help a listener whose team has a mountain of work growing beyond all proportion, and how to work with sponsors to tame (or even better, delete!) a 5-year backlog.
Squirrel: Welcome back to Troubleshooting Agile. Hi there, Jeffrey.
Jeffrey: Hi Squirrel. I’m really excited because we have a listener question! I can’t wait to share it with our with our listeners.
Squirrel: It is from an anonymous person who says: ‘We have a large backlog of portfolio items, the next six items are T-shirt sizing, and might take 9-12 months based on similar work. Beyond that is about five years of work. We are forced to size things to help with planning, but we cannot get senior staff to stop committing to work. The slower things go, the more interventions they make trying to make things go faster. I have been told to outsource some work to speed things up, but this doesn’t seem possible without diverting our internal team completely. Second, the team were forced to cut corners to meet a deadline they did not commit to, and this has caused a huge amount of technical debt. The debt never gets dealt with because of the first problem, but slows things down further. I’m stuck between a rock and a hard place. What can I do to break out of these bad patterns?’
Jeffrey: This is classic. There’s so much in here. We get The Mythical Man-Month and how adding people to late projects makes them later. We get the feedback loop between technical debt and portfolio items. People keep saying yes to things, so many good things here.
Squirrel: Good in the sense that there’s lots to work on and improve.
Jeffrey: So many great problems! I think a lot of people will recognise these because they’re pretty common as a set. ‘We keep being told to do more. That means we don’t ever have time to make things better. Things slow down. So then we get interventions imposed upon us, which makes things worse.’
Squirrel: Some of our listeners might be saying, ‘wait a minute, did I write to them?’ Don’t worry, it wasn’t you.
Jeffrey: One interesting thing here is this question, ‘what can I do to break out of these bad patterns?’ I think we can answer a couple of different ways. There’s what you could do to try to bring interventions. There’s a second thing, which is what could the organisation do to break out of the patterns. They’re slightly different, right? If we were going in there-
Squirrel: This is perfect for either of us to be consulting. So hire us! Jokingly, let’s say that’s one of the solutions.
So How to Fix This?
Jeffrey: Seriously though, part of the process to come in is to investigate where people are perceiving the problem, what problem do they perceive? If we were to go talk to senior staff here, they almost certainly see a problem: they seem to see things as being too slow. But their top down perception of the world is probably very different from the bottom up view. How might the person paying for all this see it versus how our anonymous listener sees it in the trenches?
Squirrel: Indeed. Does our anonymous listener have access to the people making those decisions? We reiterate this often on the podcast: you have more access than you think you do, especially if you think you might have some solutions to these problems. Surely these senior staff who are coming along and intervening and promising things are finding that frustrating. So I imagine that if your organisation isn’t completely dysfunctional, at least paying the bills and keeping the lights on and so on, that there’s some gold owners who are very, very frustrated with this situation and would be interested in solutions. So the first kind of schema for a solution is how can you get to those senior staff? I don’t know whether our anonymous listener meets with them every day because they’re so frustrated and intervening, or if they are in a distant place in a faraway country that this person doesn’t have access to. But in any of those cases, one of the first things I’d be doing is making sure that the mechanism for an intervention goes all the way to the top and all the way to the bottom.
Jeffrey: If we could get access to those people, I think I know your first reaction if you talk to those people and are in a position to say, ‘Yeah, I have a backlog that’s five years work.’ Why don’t we start there? What’s the first bit of advice you’d have for someone with this giant backlog?
Squirrel: The thing that I’m famous for saying, of course, is to delete the backlog. It’s completely meaningless to have something that is even 9 to 12 months out. It is much more effort to maintain the backlog than it ever has value, because if there’s something so important that we are going to do it in four and a half years, someone will remind us before then. You don’t hear ‘It’s time to do that thing that I’ve been waiting four and a half years for!’ It is much more likely that thing will not be important anymore, or the needs and requirements around it will have changed. Many changes happen over the course of four years. In this case, the T-shirt sizing this team is doing is actually wasted effort. It means those senior staff people are not getting their results because the team is busy planning something that at best won’t be relevant for four years or more realistically, will never happen. Of course, just deleting the backlog is not something you could do without those senior staff being involved, which is why I started with that mechanism.
Jeffrey: You mentioned that they’ve done the estimating for 9 to 12 months of work and then they have five years beyond that. We’re talking about getting rid of both of these backlogs?
Cut and Burn
Squirrel: Absolutely, yeah.
Jeffrey: The estimates that you’ve done for work that’s 9 to 12 months out is useless. By the time you get to something that far along, the work you did to break it down is not going to be very helpful. I think our anonymous questioner here understands that, and their point is that they’re being forced to break it down.
Squirrel: Which is why you need to get to the senior people: to explain that there’s one very basic thing you could do that would immediately improve productivity and would get them their stuff faster. To stop being quite so artificially predictable—to stop trying to look predictable—and instead actually be predictable. In other words, deliver something relatively quickly.
Jeffrey: Right. I think we’re going to have to come back to the question of who the gold owner is relative to the senior staff, because I suspect this might a situation where there’s multiple people at a senior level who are making requests that are not coordinated.
Squirrel: But that’s the actual problem! I want to underline how important it is to solve the actual problem. There’s so much that you can read, I’m sure there are tons of blog posts on how to manage your five year backlog and how to groom it and all the different processes you might follow, and there’s books that will tell you on page 74 exactly how to organise your backlog grooming…Don’t do any of this stuff, because you’re not solving the problem. The problem exists somewhere else, and all you’re doing is papering over it. It would be much better and much more effective if you could actually solve the problem that someone thinks they’re solving by bringing you yet more work for you to put in to 2027.
Jeffrey: The real problem here is the lack of coordination. Senior staff becoming coordinated on what we’re going to do is critical so that we can start planning relative to what we’re actually able to accomplish; that planning is useful because it isn’t just coming up with a continual dream list, that thing that can grow forever and unbounded. Without knowing the specifics of the situation, it’s hard to know exactly what type of dynamics are at play here, but certainly what’s being described I’ve seen many places where you might get the heads of different customer accounts, for example, all making requests for projects that swamp the ability of the team to deliver.
Squirrel: Often for purposes that have nothing to do with either the organisation’s goals or actually producing something for customers! They can be territory building or they can be trying to to reserve their space. ‘Hey, if I don’t ask for something, then in 2023 when I really need something else, I won’t be able to drop this and put that in.’ There can be very perverse incentives that can arise in a situation where no one has had the difficult conversation: what you’re actually going to do and why. That conversation is probably going to involve your gold owner.
Squirrel: For listeners who might not know, this phrase is from extreme programming. There’s the gold owner and there’s the goal donor. This is a little English pun to help convey the critical insight that the people who are telling you what to do and the person who actually has the money are often different humans.
Jeffrey: Yes. The next thing in the sequence is ‘I’ve been told to outsource some work to speed things up, that doesn’t seem possible without diverting the entire team.’
Squirrel: That would be obvious. It’s just so easy, right? There’s too much work and it would take five years. So what we need is five times as many people, right?
Multiplication and Teams
Jeffrey: It does seem to be the standard answer. For people who don’t know better it does seem like an easy solution. ‘If we don’t have enough people, let’s get more. I mean, there are companies that do this now.’ There’s a problem here, which is the trade-off involved to bring those other people in: there’s time and effort spent in education. I think it’s really interesting that there are times where it would make sense to outsource, but what are the trade-offs involved? Is that something you’re actually able to do? People have outsourced maintenance of their existing systems to have their internal team go and build the next new thing, teams also do the opposite where they have a new project and outsource that. These things are possible. The challenge is to identify if that is the right thing to do. If people aren’t already coordinated, I suspect this would only make that worse, but I don’t know. What do you think, Squirrel?
Squirrel: I’m reminded of an old saw in the world of urban planning: if you build a wider road or more roads in order to resolve traffic congestion, you get the same amount of traffic congestion on the wider or more numerous roads. The reason being, people notice the change and go drive there. If we’ve got senior staff who are busy committing to new things and if we give them more capacity, they’ll just commit to more and we’ll still wind up with a five year backlog. That’s what I would predict. Even if somehow we manage to get hold of that problem and we didn’t increase the total number of commitments, we’re going to have to multiply the team by five! In a book from the 1970s that my father had on the shelf called The Mythical Man-Month I learned this is not going to work, because adding people to a late project makes it later; you have to bring them all up to speed. This is exactly what our anonymous correspondent refers to. We’re going to divert the internal team to teaching everyone? If you suddenly overwhelm them with five times as many people to try to bring it down to one year, even if you aren’t taking on new work to match the capacity, you’re going to have a counterproductive ramping up period trying to get these folks to become productive. I almost always see when I come in on consulting projects that it is much more effective to stop hiring, stop adding outsourced people, and stop adding more complexity. The team isn’t managing the current complexity very well, and it would be much better to figure out how to manage complexity and how to deal with these issues before attempting to add people. That’s when the types of interventions you described make sense. When the team is healty, you can move the maintenance out or move the new projects out or split it in some other way. But adding more people to a dysfunctional team does not make the team functional, it only makes the problem worse.
Jeffrey: That principle of urban planning is called induced demand. There’s a bit of The Mythical Man-Month there, right? In order to address congestion they decide to add a lane, but in the process of adding the lane, what do they often do? They close a lane while they’re working! You lose capacity in a way that’s actually remarkably similar to what might happen with outsourcing. You have this period where you’ll be training people up and you actually have less capacity. If they’re not managing things well currently, they’re not in a shape to be expanding. You want to have a functional system and then you can add people. Adding people to a system that’s not working will definitely make it worse because there’s also the added complexity of the additional communication overhead. This is famously an N-squared problem: the more people you have the more lines of communication there are and the more overhead. So I agree with you that this seems like an approach I would not recommend.
Squirrel: It is so important that we make this a shared problem. When I come in as an outside consultant it’s easier for me to make it a shared problem because I can say things like ‘this team isn’t very mature at handling complex demands. We aren’t very good at saying no. The backlog is ridiculously long and we haven’t done anything about it.’ It’s easier for me to say those things as an outside person and to make them shared problems with those senior staff and gold owners. Once that’s a shared problem, then it becomes much easier to say, ‘Well, one of our solutions could be to close a lane or the whole road and have less capacity for a while and then add capacity. Is that something you’d like?’ I almost never hear ‘Yes, I’d like that.’ That’s not generally where people want to go. They want to continue using the road and putting more traffic down it, but they don’t want to stop for six months. There’s rare exceptions, but that’s not typically what people want to do. But for everyone to make a much more informed decision, somebody has to start with ‘Hey, wait, the emperor has no clothes. Our backlog is five years long. There’s something wrong here, and this is a problem for our organisation, not just for technology.’
On the Same Side
Jeffrey: One thing that stood out to me is you started that by saying ‘this team isn’t handling complexity well.’ Crucially, by ‘the team’ you were including not just the development team, but all the senior staff as well.
Squirrel: Absolutely, yes.
Jeffrey: This is to say that you are a team! We say that teams are defined by people who share a problem, and this is a shared problem. Getting alignment of what the problem is and what your options are is really key. When we come in as consultants, what we’re doing is looking at it like this, we’re defining that shared problem amongst the group. It may seem obvious to you what the shared problem is, and therefore it’s harder to bring up because, well, it’s obvious, right? But in fact these things are worth stating, which might be threatening. That’s the ‘difficult conversation’ aspect.
Squirrel: Exactly. It’s making that topic discussable. You can make it discussable that we as an organisation are dealing with complexity poorly and therefore adding more complexity in the form of additional people in Remote-istan might be a poor idea. Then you’re able to have that discussion much more effectively, but it’s often left unstated because it’s threatening.
Jeffrey: In part because if I’m in the position of our listener, I want to go have this conversation. But, I know I’m going have to go in and say to people, ‘I bet you’re frustrated these things are taking so long. I bet you’re unhappy with us becasue we can’t finish these as quickly as you like and in fact I’m telling you it’s going to be nine months. We’ve broken down, and you’re probably really unhappy to hear that it’s going to be five years for all this other stuff. Am I right about that?’ Asking those kind of questions and being vulnerable in that way is going to be very challenging.
Squirrel: On the other hand, once you make it discussable, you can actually do something about it.
Jeffrey: Exactly. So for this first type of problem here about the portfolio, it’s definitely about this difficult conversations to build alignment with the team and including the gold owner. I think it probably makes sense for us to to take on part two next week, because I think they are different in character. This first one is very much about external co-ordination and the second one is about cutting corners, which is going to be a bit different.
Squirrel: Good. So we’ll have some interesting topics to cover next time. Thanks, Jeffrey. Thanks, Squirrel.
Jeffrey: Thanks, Squirrel.