TBM 341: 10 Tips for Turning Around a Platform Team
Platform teams are tough work. The idea sounds great, but the day-to-day is hard. Someone recently reached out for advice with this common scenario. If you are a platform team member, you might find this helpful. Otherwise, consider forwarding it if you know a team in this situation.
Scenario:
A platform team is underwater, and things keep getting worse.
Like most platform teams, they must balance advancing their roadmap against the ad hoc requests from other teams. When (if) the platform fully materializes, these requests will be (more) self-service for requesting teams, but the work is manual for now. Handling the requests does provide some opportunities to learn about partner needs, but the context-switching is very distracting. They're losing ground—two steps forward, three steps back.
The platform team reports to a different group from the requesting teams, and you have to go a fair way up the org chart for any "tie-breaking." Every day feels like a juggling act without a North Star. When leaders originally green-lighted the platform idea, they didn't think through the implications of prioritization and their respective goals.
The team has tried putting on the brakes when it comes to outside requests, but partner teams work around them, further complicating future efforts to converge on platform-like solutions. The partner teams lose confidence in the group. Meanwhile, their direct managers are perpetually frustrated by slow progress on the platform work. There's a wicked cycle whereby they keep getting asked for commitments and reply with commitments, and those commitments are seen as "not enough," so they fall into overcommitting on their roadmap and fall short.
My Reply
None of this is easy, but there are some things you can do to increase the odds of wiggling out of this tough situation.
Stop talking about debt and abstract concepts. Platform talk can be very heady. I enjoy it. But most people see platform talk as an endless theoretical discussion that sets back your cause. Keep things very simple, and don't say things that seem like random bits of complaining or theorizing. Keep your platform promise—the intended impact and how it will be a force multiplier for the org—front and center. Use every opportunity to talk about progress, blockers, and what you intend to do to address those blockers. Be a broken record.
Stop shooting yourselves in the foot. I know this is hard to hear, but you must stop doing things that perpetuate the cycle. Approaching each quarter with rose-colored glasses might temporarily keep managers off your back, but you'll end up in an even worse situation. Playing Tetris with your capacity will ultimately backfire. You've already lost the game if you're thinking about five percentage point allocation units and a twelve-week (or more) timeframe. Even if you need to play external optics games externally, internally within the team, you'll need to be extremely judicious about scope, limiting WIP, doing the highest leverage work possible, working small, and vetting external requests. I see so many teams make an already hard problem even worse.
Related, you have to stop optimizing for being underwater. Lots of teams in this situation become more fractured. They try to behave like five teams of one instead of one team of five, and they send off tech leads, engineering managers, and product managers to proxy stuff and play defense instead of focusing on the task at hand. If you behave like you're putting out a million fires all the time, that is what you'll do.
Make aggressive use of enabling constraints. You must become a super-user of enabling constraints to wiggle out of this. For example, only think about external requests every other Friday, create a max batch size limit of 3 days of work, limit team WIP to [Team members] * 0.5, pair all day Tuesday, etc. Put it this way, somehow / someway behavior needs to change. Otherwise, you will repeat past cycles. To change behavior, you need to commit to something being very different. Nothing will change if you keep getting mired in trivial trade-offs while the train is literally leaving the station. Tiny little micro-adjustments here and there will not cut it. What do you have to lose? Go big on enabling constraints.
Encourage contributions. Take it if you can get anyone to contribute their time and work to your cause! It might not be perfect, and your team may need to assist, but take any help or contributions you receive. Build an extended network of friendly contributors across the company.
Try to get your managers and the partner team managers in a room together whenever possible. Talking to one party or the other will never highlight the issue. If you can get both parties in the room, you can highlight that their lack of alignment is one of the root issues here. But you have to do this with the utmost care and political savvy—framing it as a way for both your managers and your partner managers to get what they want.
Keep receipts. I'm sorry, but you will need to track your time inside the team so that you can confidently walk into meetings and give a play-by-play of the current impact of requests, etc. Normally, time-tracking is a big waste of time, but when your team is getting so much scrutiny, and there are all sorts of pet theories about what is happening, you need to commit internally to being hyper-aware of where the time and energy go.
Full transparency. Keep your queues transparent at all times. Where are the requests coming from? How long have they been in the queue? What is your SLA for getting back to people? What are the cycle times and lead times for requests? How does that impact the time and energy on your platform roadmap?
Work very small on your platform roadmap. I know some things take months and quarters, but you must be extremely meticulous about taking thin slices of your platform roadmap and getting those things out the door. Be aggressive when defining milestones and stepping stones to represent meaningful pivot points in your efforts. Don't just work down a burndown on auto-pilot.
Figure out how to keep the team motivated. Wiggling out of a situation like this could take a couple of quarters or years. It is easy for a team to get demotivated. They need to see progress, but the right kind of progress. Progress, in this case, might be the number of external requests that can be coaxed into a self-service motion, contributions from other teams, or a percentage of the week dedicated to focused work. Whatever motivates the team, keep that front and center.
At a high level, this is the classic "fixing the train while the train is barrelling down the tracks" problem. Once you lean into that instead of looking for quick fixes like freezes, big-bang re-plans, hail mary projects, etc., your team will embrace the rigor and discipline it will require to shift the tide.
Above all, remember that slipping into the optics game and planning traps for short-term wins, almost always backfire.
Hope this helps!
As an ex-Engineering manager of a platform team, this is gold. Thank you.
Thanks for this post!
One addition is to learn about funding models. This may theoretically be “above your pay grade” as a line-manager, but in a platform team you have no choice.
Start by figuring out where your headcount (HC) is coming from. Who is paying your salary? You will often be surprised how much fragmented history you can uncover digging this up.
Study up on the different funding models. This can be “lending HC,” internal budget moves, something like the Amazon “Away Team” model, and more. Each has pros and cons, so make sure to understand the best practices for each model.
Then, you have to work with your leadership to get buy-in for a SMALL set of funding models that they will agree to. Solicit feedback from your stakeholders and leadership to get to one *published* document with the various acceptable options. Agree on an escalation path if none of the models work for a given request.