Stop coaching from the best call of the week

In one week of scored calls, roughly three in four agents produced at least one call in their organisation's top 10%. The best call of the week is a good day, not a good habit. It shows you what someone can do when everything lines up, and almost nothing about what they do the rest of the time. A team told to copy it copies the day.

The case for playing the best call

I want to argue the other side properly first, because it's a good argument.

A great call is concrete. Coaching advice is abstract. "Ask better discovery questions" has never once changed what a rep says at 10:14 on a Tuesday, and everyone in the room knows it. Play 40 seconds of someone actually asking the question, the pause after it, the customer filling the silence, and the room goes quiet. People copy what they hear. They mostly ignore what they're told.

A manager who plays a good call on Monday is doing more than most. The alternative in a lot of teams is a slide about the funnel. (Same slide since 2019, by the look of it.)

So I'm not telling you to stop playing calls. I'm telling you the selection rule is broken, and one number changed my mind on that.

Three in four agents had a top call last week

We score every call, which is the only reason the ordinary ones have numbers at all. When I pulled a week of them, roughly three in four agents had at least one call in their own organisation's top 10%. Not a handful of stars hoarding the good work. Most of the team.

So "who had the best call this week" pulls one ticket out of a pile three quarters of your people are already holding.

Now do the sums on the cadence. You play one call a week and never miss a week, so you get through 52 calls in a year. The top 10% of a single week already runs to hundreds. Your entire year of examples is smaller than the good work one week threw off, and you chose it by whichever call happened to score highest on the day. (Call libraries have the same problem, just with better search.)

What the best call says about the person

Here's the number.

Take agents with at least 10 scored calls in the week, so we're looking at people who spend the day on the phone. Their best call gets to about three quarters of what their scorecard asks for. Their middling call that same week, same person, same customers, same script, gets to about six in ten.

Then compare agents to each other. Their ordinary calls sit far closer together than that. The distance between one rep's best day and their normal day is more than twice the distance between one rep and the next.

Read that with a manager's hat on. The gap you rank people on is smaller than the gap inside the person you're about to hold up as the example. You'd learn more comparing a rep to themselves than to the leaderboard, and it isn't close.

Picking the top of a noisy pile

A call is one conversation, with one customer, who was in one mood. Some of the score is the rep. Some of it is the day. Take the highest one and you've selected for both without knowing the split.

That's survivorship bias, which here just means you're staring at the winners and can't see the pile they climbed out of. It's also why the follow-up call comes back to earth, which has a name of its own: regression to the mean. A rep who blows the doors off on Tuesday hasn't gone backwards on Wednesday. They're back where they live.

The best calls stop short too

The call you play as the model is missing about a quarter of what you asked for. Not one call in the week's top tier got everything the scorecard wanted. Not one. The tier averages about three quarters, and virtually all of it sits below 90%.

The week's best calls stop short of the scorecard

Every scored call in one week, grouped by how much of its scorecard it got to.

The right-hand end of that chart is empty. Fine as an example of one thing. Rough as a template.

There's a catch: this is one week of calls, from teams that score every one of them, so read it as a snapshot of how those teams run rather than a trend or an industry picture.

What I'd do on Monday instead

Pick one behaviour you want the team better at this month. Asking for the next step and then staying quiet, say. Then look at how that single behaviour scored across every call in the week, the ugly ones included, rather than at which call scored highest overall. You'll usually find a perfectly ordinary call that nailed it, and that one is worth more in the meeting, because everything else about it looks like a normal Tuesday.

Compare each rep to their own ordinary week. That's where the wider gap sits, so that's where coaching has something to grab.

If you do play a call, say what it's an example of before you press play. One behaviour, named out loud. Then be willing to play a middling call that did that one thing beautifully. I've argued separately about how often to run these reviews; this is about which calls you pull when you do.

CallCoach is how we score every call, the ordinary Tuesdays included. Try it on your own calls and see whether your best call is a good day or a good habit.

---

All CallCoach data in this post is aggregated, anonymised performance data from hundreds of thousands of calls across a variety of industries. We report the trends we spot; full statistical analysis is beyond the scope of a blog post.