Working Too Smart
A diagnostic for when a good playbook stops fitting the account — and what to do next. A good sequence or signal handles most accounts; it won't handle the one where the customer isn't responding the way the play assumes. Use it to spot when a play has stopped fitting the account in front of you, decide whether to override it, and check yourself account by account.
Reps are accountable for the outcome, not for completing the sequence. When the play stops moving the account, the job isn't finished.
"I ran them through the sequence."
Listen for it this weekA manager asks how an account is going; the rep says they ran the sequence. Did it land? No — and the conversation stops there, as if a play failing were the end of the work rather than the start of it. This is common on teams that have invested heavily in process: once work is measured by what the workflow says you did, it's easy to stop asking what this particular customer needs.
The Process Veto
"We can't work that account, there's no email in the CRM." A data-hygiene rule is now deciding who gets called — a choice no one actually made or noticed.
The Cadence Ceiling
"We don't want to call them too much." Some customers only pick up on the fifth try in a single day. The cadence has capped effort below what the account needs.
The Completed Task
The steps are green and the CRM is clean, but the customer still isn't using the product. Activity got reported as an outcome, and no one checked whether any of it landed.
The Comfortable Silence
An automated nudge is easier than a hard conversation, so no one tells the customer they've done none of the setup and the launch is at risk.
A dashboard shows that an account is at risk; finding out why is still the person's job.
Signal, Play, Override
The system surfaces the signal. The person decides what this customer needs. It's easy to instrument the signal and the play well and leave the third step — the override — undefined. When that happens, the play ends, the account goes quiet, and there's no defined move out of "no."
The override step is the one most often left undefined. Without it, the sequence expires, the account goes quiet, and every task still shows complete.
What an override looks like
An override is a deliberate departure from the default, made by the person closest to the customer, without waiting for permission. Four conditions justify one.
- The play ran to completion and nothing about the customer's behavior changed.
- The signal contradicts the play: the data says you lost your champion, the sequence says feature training.
- The account is worth more than the cost of breaking the process.
- The customer already told you how they want to be reached, and it isn't in the sequence.
e.g., twelve touches over about three weeks, one template set for the segment.
e.g., seven calls over four to six weeks, homework between each one.
Health-score alert, quarterly check-in, nurture email, then a save sequence.
Reply, resolve, close, send the CSAT survey.
Judgment needs support from the leader
You can't ask a team to work outside the workflow and then grade them only on the workflow. Four responsibilities sit with the leader, not the rep.
Ship fewer signals
Too many signals train the team to ignore the dashboard. Keep the set small — a suggested three to five per account — each tied to a churn reason you're actively working on.
Publish plays as suggestions
Call it a suggested flow, not a script. Write down what everyone already knows: it works most of the time, and here's how you'll know when it isn't.
Grade the outcome, track the tool
Don't let "I ran the sequence" end a conversation. Track each play's hit rate yourself — that's the leader's job. A play performing below its bar gets rebuilt or retired, not enforced harder.
Grant permission out loud
Including permission to risk the account. A coach who can't say "this implementation is failing; let's stop" tends to lose the account slowly instead of saving it quickly.
Coach it in these words
The change is largely verbal: change the first question you ask about an account. Changing that question is a lever leaders can pull directly — treat the time it takes behavior to follow as something to observe, not a promise.
Better first questions
+ "Did it land?" before "did you do it?"
+ "This flow is suggested — you own the outcome"
+ "What did they actually say when you asked why?"
+ "What are you going to try that isn't in the playbook?"
+ "That save was creative — tell the team how you did it."
Habits to retire
– Treating a finished sequence as the end of what's possible
– Capping contact out of a vague fear of "too often"
– Skipping an account because the data isn't clean
– Enrolling a stalled account in the next sequence by default
– Handing reps a long, undifferentiated list of signals
Software can flag that a sequence isn't working; deciding what a specific customer needs instead is still a person's judgment.
The thirty-second check, per account
Run this per account in your one-on-ones. Any unchecked box is the next thing to do by hand — not the next thing to automate.
- I know whether my last action landed, and I'm not guessing.
- I can say why this customer isn't moving in their own words, not in dashboard categories.
- I have said the uncomfortable thing to them out loud, at least once.
- I have done one thing for this account this week that no playbook told me to do.
- If every automation switched off tomorrow, this account would still be handled.