The problem in numbers
Pull last month's leads and compare two timestamps: lead created and first activity marked done. Most teams that measure this for the first time find a median in the range of one to three days, with a long tail of leads that waited a week. Meanwhile the research on lead response has said the same thing for years: the odds of reaching and qualifying a lead collapse within hours of the enquiry. Buyers shortlist whoever answers first, and your proposal quality never gets a chance to matter.
The gap persists because nobody owns it. Response time sits between marketing, who count the lead as delivered, and sales, who count the deal from first conversation. An SLA makes the in-between measurable and owned. That is all it is: a target, a timestamp pair and a consequence.
The design
Define the clock precisely, in writing. Start: lead created in Pipedrive. Stop: first outbound contact attempt logged, meaning a call activity marked done or a tracked email sent from the lead. Define business hours, because a Friday 17:55 lead answered Monday at 09:15 met a one-hour business SLA. Vague clocks produce arguments instead of improvements.
Then set two or three tiers. A practical SMB set: demo requests and inbound calls at one business hour, other inbound leads at four business hours, imports and list uploads at two business days. Each tier gets a Response tier single option field on the lead, set by the same rules that route it. The tier travels with the lead, so a rep never has to remember which target applies.
Enforcement runs on three light automations. On assignment, create a follow-up activity due inside the SLA window, named for the tier, so the target lands in the rep's task list. At the SLA boundary, escalate quietly: a filter of leads past due with no done activity, delivered to the sales manager as a daily view or Slack push. And keep the routing notification instant, because a rep cannot respond to a lead they have not seen. Note what is missing: no automation nags the rep directly at minute fifty. The escalation goes to the person who manages capacity, because a missed SLA is usually a staffing signal, not a laziness signal.
Implementation steps
Step one, measure the baseline before announcing anything. Two weeks of silent data beats any assumption and sets the target conversation on facts. Step two, agree the tiers with the team, using the baseline. Step three, add the tier field and wire it into the routing rules. Step four, build the follow-up activity automation and the overdue filter, following the restraint principles in automations that matter. Step five, put the SLA numbers into the weekly sales meeting, next to pipeline and forecast, for at least a quarter. What leadership looks at weekly, reps optimise daily.
How you measure that it works
Two numbers per tier, monthly: median time to first contact, and percentage inside SLA. Build both in Insights from activity reports, filtered per response tier. Watch the compliance number, not just the median. A team can hold a fine median while a fifth of leads quietly miss the window, and that fifth is where the churn of trust happens. Target eighty percent compliance at the start. Above ninety-five, consider tightening the target. Persistent misses in one tier mean the tier is mis-staffed, not that the team is failing.
Common failure modes
Gaming by activity: reps mark a token activity done to stop the clock without real contact. Counter it by spot-checking escalated leads and by defining the stop as a call or sent email, not any activity type. Target inflation: the SLA set at the best rep's pace, missed by everyone else, then ignored by everyone. Set targets at the eightieth percentile and tighten quarterly. Nag fatigue: automations that ping the rep at every threshold train the team to filter notifications. One notification at assignment, one escalation to the manager, nothing else. And clock rot: business hours change, holiday calendars differ, and the measurement silently drifts. Review the clock definition twice a year alongside your hygiene routine.