Speed to lead is the time between someone raising their hand and someone from your team actually speaking to them. Not the automated email. Not the SMS. The conversation. It is the least discussed number in most sales organisations and routinely the most decisive. What happens inside that conversation is the other half, and it decides sales call conversion.
Why does speed to lead matter more than messaging?
Intent decays fast, and it decays for a mundane reason: the person who enquired is doing something else now. They filled in a form between meetings, or called three clinics in a row on their lunch break. The one who calls back while that context is still loaded gets a conversation. The one who calls back tomorrow gets a confused person who has already booked elsewhere.
This is also why 'we contacted them within 24 hours' is a meaningless SLA. It measures compliance with a process rather than presence at the moment of intent.
The mechanism is competitive, not psychological. Someone enquiring about a service has usually enquired more than once, and the first real conversation frames every comparison that follows. You are not persuading a patient buyer; you are arriving before or after the person who took the decision away from you.
How do you measure speed to lead properly?
Most teams that report a good response time are measuring the wrong event. Three common errors:
- Counting the automated acknowledgement as first contact. It isn't: nobody has spoken to anyone.
- Counting the first dial attempt rather than the first connected conversation. An unanswered dial is not a response.
- Reporting the mean. One lead contacted in nine seconds and one in nine hours average out to something respectable and describe nothing. Use the median and the 90th percentile.
The 90th percentile is the one worth putting on a wall. Median speed to lead tells you what a typical enquiry experiences; the 90th tells you how bad the tail is, and the tail is where the lost revenue sits. A team with a five-minute median and a six-hour 90th has a routing problem affecting one lead in ten, and the median will never show it.
The three fixes, in order of payoff
- 1Remove the routing delay. Most of the lag isn't reps being slow; it's a lead sitting in a queue, a sync running every fifteen minutes, or an assignment rule waiting for a manager. Instrument the handoff before you coach anybody. Reps cannot beat a fifteen-minute sync no matter how motivated they are.
- 2Rank the queue by intent instead of by arrival. Not all enquiries deserve the same urgency, and first-in-first-out guarantees your best lead waits behind three tyre-kickers.
- 3Close the loop on unreturned enquiries. Every organisation has leads that were assigned and never actually called. Nobody knows which ones, because there's no report that joins 'lead created' to 'conversation happened'. Callix builds that join from the calls.
Work the three fixes in that order, because they are ordered by what they cost you. Removing routing delay is a configuration change nobody has to be persuaded of. Ranking the queue is a decision about what counts as intent, which takes a meeting. Closing the loop on unreturned enquiries is the one that requires a report nobody currently has, joining the moment a lead was created to the moment a conversation actually happened, and it is also the one that recovers revenue you have already paid for.
A note on what not to do with the number. Publishing a per-rep response-time leaderboard reliably produces exactly what you measured: fast, useless dials that connect to nobody and close the SLA. If you report it individually at all, report connected conversations rather than attempts, and expect the behaviour to follow the definition rather than the intent behind it.
What does good look like in practice?
There is no universal target, but the shape of a healthy number is consistent. Contact in single-digit minutes for enquiries arriving during working hours, a 90th percentile inside the hour, and a share of never-contacted enquiries close to zero. If any of the three is missing, the other two flatter you.
Get there by fixing the system rather than the people. Every minute of the delay belongs to something specific: a form posting to a queue, a sync interval, an assignment rule, a rep in another call. Attribute the minutes before you set a target, or the target lands on whoever is easiest to hold accountable rather than on whatever is actually slow.
When should you stop optimising response time?
Speed to lead has a floor. Once you're consistently reaching people in minutes, further gains are marginal and the constraint moves to what happens in the conversation itself: whether an offer was made, whether the objection was answered, whether a next step was agreed. Teams that keep grinding on response time past that point are optimising a solved problem while the real one goes unmeasured.
There is one exception worth naming. If your enquiries arrive overnight or at weekends, the floor is set by coverage rather than by process, and no amount of queue tuning moves it. That is a staffing or automation decision, and it should be argued with the hourly distribution of enquiries in hand, which differs sharply by the kind of operation you run, rather than as a general principle about being fast.