How to deal with low performers on a mob programming team?

I’ve talked about mob programming before, but never about how to handle under-performers on mobbing teams.

What does under-performance look like on a mobbing team?

Low performers usually fit one of these two molds:

[More common] Fly On The Wall Felix – He coasts along in read-only mode. He’s on the mobs, but he doesn’t contribute much and doesn’t appear engaged. He may as well not be there at all. And if you’re thinking “but some people are just quiet!” then that’s different. Some of my favorite people to mob with are silent, but attentive. They pay close attention and catch things others miss, and they contribute relevant links and context in the Zoom chat rather than speaking it.

[Less common] Distracting Dory (like from Finding Nemo? Get it?) – She distracts the team with tangents, rabbit holes, repeated questions, or irrelevant information. She’s engaged and active, but she hurts velocity and frustrates everyone, so it would be better if she wasn’t there. Don’t give up on Dory; she’s probably coachable! That energy is great, but it needs to be focused by raising her “is this important enough to say?” threshold.

And of course, on any team (mobbing or not), bad apples are under-performing. Downers and jerks are always unacceptable regardless of what else they bring.

How can a manager identify low performers on a mobbing team?

If the manager joins the mobs every now and then, it’ll probably be obvious. If someone almost never contributes anything, they’re probably a Felix, and if someone seems to annoy the team whenever they speak, they’re probably a Dory. Track it to see if the pattern continues over a few sessions, because everyone has a bad day or gets into a weird mood sometimes.

If the manager doesn’t join the mobs, or if the low performer puts on a show when the manager joins, then the manager has to rely on the team to tell them. This depends on three things:

• High psychological safety, so the team feels comfortable telling the manager
• High standards and accountability, so the team isn’t willing to tolerate it
• A challenging problem or deadline, so the team feels the pain of low performance

Once the low performer has been identified, then they have to do something about it, which brings us to…

How can managers coach low performers on a mobbing team?

It’s no different from coaching anyone on any performance issue.

• Step 1: Give the feedback. Something like: “When you don’t contribute on mobbing sessions, I feel worried that the team isn’t benefitting from your skills, knowledge, and expertise.”
• Step 2: Agree on a plan. I like the GROW model for this. That way the low performer can take an active role in defining the problem (Goal vs. Reality) and the plan (Options/Obstacles and Way Forward).
• Step 3: Track it. You don’t need super accurate or detailed metrics at this stage. Drop into a few mobbing sessions and see for yourself how it’s going, and ask the other teammates if they’ve noticed any change/improvement.

How can managers fire low performers on a mobbing team?

If performance doesn’t improve, then it’s the manager’s job to get them off the team.

Firing people is so hard you’ll invent excuses not to do it, and this is especially true on a mobbing team. It’s easy to rationalize with “well the mobbing sessions are still productive” and “maybe it’ll fix itself when they get more comfortable with leading mobs.” No! Give up earlier! Stay firm and harden your heart.

But it’s tricky. At many (most?) companies, a manager can’t just decide to fire someone. HR requires managers to show evidence of consistently not meeting expectations for multiple months before any firing can happen. Some companies require a PIP and some don’t, but either way, the manager has to point at data. Data, data, data.

How can you get data on individual performance from a mobbing team where people do things together? Here are some possible data sources to mine:

Feedback from teammates. Have everyone rate each mobbing session, and compare the ratings of sessions led by the low performer vs. others. Or do some good old fashioned peer feedback sessions, even if it’s not review time. I like these questions but the specific questions don’t matter here.

Mobbing sessions driven. Especially for Felixes, increasing this number of mobbing sessions they lead is a good signal of increasing engagement. It’s pretty hard to be a fly on the wall when you’re the one driving.

Contributions per session. This could be chat comments, spoken suggestions, on-point questions, relevant links found, whatever. For Felixes, you want a raw count, and for Dorys, you want to track “helpful vs. unhelpful” contributions.

PRs and issue activity. Usually on a mobbing session, the person driving or the person who was especially impactful will “get the honor” of creating the PR or leaving the comment. So PRs created or closed, issues opened or commented or closed, all of that is still relevant, and particularly hard to game on a mobbing team.

Self-described impact. Have them DM you the answer to this question every day: “What did I do to make an impact today?” The expectation is that 1) their answers are true and can be verified and 2) their impact becomes greater over time. If either of those aren’t the case, that’s important evidence. This also protects against firing someone who has invisible but important impact.

Those plus any other team specific metrics plus a dash of creativity should be enough to build sufficient evidence that someone isn’t performing and needs to be terminated.