Enquiries are steady, the ad spend hasn't changed, and yet fewer deals close than last quarter. The usual suspects are the campaigns or the sales script. Quite often, though, the money leaks out earlier — in the gap between "the customer got in touch" and "someone actually replied". Few teams measure that gap at all, and those that do tend to measure it in a way that produces a comfortable number and hides the real problem.
This guide covers how to measure speed to lead honestly: what counts as a first response, which data you need, why the average misleads, how to slice the numbers to find the cause, and what target to set for a sales team. At the end, we look at where AI genuinely helps and where it does not.
Why this metric matters more than most
An enquiry cools quickly. At the moment someone fills in your form or sends a WhatsApp message, they are thinking about your service. An hour later they are in a meeting; by the evening they have contacted three other providers. Whoever replies first with something useful gets to frame the conversation — what a fair price looks like, what timelines are realistic, which criteria matter.
There is research behind this, too. In "The Short Life of Online Sales Leads" (Harvard Business Review, 2011), the authors found that companies that tried to contact a lead within an hour were nearly seven times as likely to have a meaningful conversation with a decision-maker as those that waited longer. The data comes from US companies over a decade ago, so do not copy the percentages into your own plan. The direction, however, is clear — and easy to test on your own numbers.
The second reason is control. Conversion to a deal depends on dozens of factors: product, price, season, traffic quality. Response time depends almost entirely on how work is organised inside your business. It is one of the few sales metrics you can move within a week without touching the marketing budget.
Defining a "first response"
This is where most measurements fall apart. Unless the definitions are agreed in advance, every salesperson will count differently, and the numbers will not be comparable.
- The clock starts when the enquiry lands anywhere in the business: a website form, a WhatsApp or Telegram message, an Instagram DM, a missed call. Not when someone "saw" it or moved it to "in progress" in the CRM.
- The clock stops at the first substantive reply: an answer to the question, a clarifying question, a proposed call time, or a connected call.
- Not a response: an auto-reply such as "Thanks, we'll be in touch shortly", a status change in the CRM, or an unanswered call with no follow-up message. Auto-replies are useful reassurance, but the customer learns nothing new from them.
Agree separately how to handle out-of-hours enquiries. It helps to track two numbers: response time during business hours (a measure of team discipline) and response time across all hours (what the customer actually experiences). This matters a lot for businesses in the Gulf, where the working week, prayer times and Ramadan hours shift the schedule, and for UK firms that get evening enquiries from people browsing after work. The gap between the two numbers is usually the strongest argument for change.
Where to get the data
The ideal setup is a CRM that receives every channel automatically, with timestamps on every inbound and outbound message. Then the metric is a single query. In reality, the website feeds the CRM, WhatsApp and Instagram live on salespeople's phones, and calls sit in a separate phone system. In that case there are three practical routes:
- Exports. From the CRM: deal creation time and the first outbound activity. From telephony: missed calls and callbacks. From messaging apps: chat exports. Combining these into a spreadsheet by hand for one or two weeks is entirely doable.
- Mystery shopping. Send 15–20 enquiries through different channels at different times — including evenings and weekends — and record when a reply came and what it said. It is not statistics, but it is vivid, especially for an owner who believes the team "always replies fast".
- Bringing channels together. Connect messaging apps and forms to one place where every message has a timestamp. This is a process change rather than a measurement, but without it the metric will never calculate itself.
We walk through connecting messengers and forms into a single funnel in "AI in sales: automating the funnel from lead to deal".
Why the average lies
Say ten enquiries arrive in a day. Nine get a reply within 3 minutes; one gets a reply 14 hours later because it came in at 10 pm. The average is roughly 87 minutes. That reads as "the team replies in about an hour and a half", when in fact nine customers were served brilliantly and one was almost certainly lost. The opposite happens too: a handful of instant replies can mask a dozen enquiries that sat untouched for half a day.
So look at the distribution, not the mean:
- Median — the typical customer experience.
- 90th percentile — how long the unlucky ones wait. This is where lost enquiries live.
- Share within target — for example, "what percentage of enquiries got a substantive reply within 15 minutes". This is the easiest number to explain to a team.
- Share never answered — enquiries that nobody replied to within 24 hours. It is usually invisible until you start counting.
Breakdowns that reveal the cause
The headline number tells you there is a problem. The breakdowns tell you where it is. A minimum set:
| Breakdown | What it typically reveals | What to do about it |
|---|---|---|
| By channel | Web forms are handled quickly; WhatsApp and Instagram DMs have long gaps | Bring channels into one inbox and give each an owner |
| By hour and weekday | Lunch, evenings and weekends are black holes | Rotas, shifted hours, or an automated substantive first reply |
| By salesperson | Several-fold differences between people | Look at causes: workload, lead distribution, habits |
| By enquiry type | Simple questions wait as long as complex ones | Hand routine answers to a knowledge base or AI |
| By volume | On peak days response time multiplies | Redistribution, a queue, an automated first reply |
The last breakdown is especially telling: if response time climbs after a campaign launches, some of the enquiries you paid for are lost at the door. To estimate what that costs, use our lost-revenue calculator with your own figures.
Setting a target
There is no universal number. A beauty salon and an industrial equipment supplier need different speeds, and a B2B buyer in London has different expectations from a consumer messaging a Dubai clinic at midnight. A sensible sequence:
- Measure for two weeks without changing anything. Get the median and 90th percentile by channel.
- Compare conversion to the next stage for enquiries answered quickly versus slowly. This is your own version of the HBR study, and it will persuade your team more than anyone else's data.
- Set a target the team can genuinely sustain during working hours — for example, "90% of enquiries get a substantive reply within 15 minutes".
- Decide separately what happens outside working hours: an on-call person, a first-thing-in-the-morning priority, or an automated first reply.
- Put the metric where people see it daily: a dashboard or a morning summary in the team chat. A metric reviewed once a month does not change behaviour.
Where AI helps — and where it doesn't
An AI assistant covers the most expensive part: the first minutes and the hours when nobody is at a desk. It replies immediately, in the customer's language, asks clarifying questions and hands the salesperson a qualified request rather than a bare "new lead". For a walkthrough using Instagram as the example, including where such setups break, see "An AI assistant for Instagram DMs".
| Stage | What AI does | What stays with people |
|---|---|---|
| First reply | Answers within seconds, at any hour | Checking tone and rules at launch |
| Routine questions | Timelines, location, terms — from the knowledge base | Keeping the knowledge base current |
| Qualification | Clarifies the need, collects contact details | Deciding whether to take the customer on |
| Handover | Creates a CRM deal with the full conversation | The call, negotiation, pricing, closing |
| Measurement | Timestamps every message | Reviewing the numbers and changing the process |
The honest limits: if a salesperson picks up the conversation half a day after the AI's quick reply, the customer still cools — just later. So alongside the assistant, measure a second interval: from handover to the first human action. And AI cannot fix a business that has no clear answers to its own frequent questions; it will have nothing useful to say.
More on how these assistants are built is on our AI agents for business page. Other articles on the topic are in the Sales section.
FAQ
Does an auto-reply like "we'll be in touch soon" count as a first response?
No. It tells the customer the message arrived, but gives them nothing of substance. Track it separately if you like, but keep it out of the metric — otherwise the number drops to almost zero and stops being useful.
What about enquiries that arrive overnight?
Track two numbers: business hours and all hours. If overnight enquiries are rare, a "first thing in the morning" rule may be enough. If there are many — common for consumer businesses in the Gulf, where evenings are peak browsing time — you need an on-call person or an automated substantive first reply.
How much data do I need before drawing conclusions?
Two weeks and a few dozen enquiries are enough for a first picture. Conclusions about individual salespeople or specific hours need more: on small samples, one odd enquiry shifts the number a lot.
Won't salespeople just reply fast but say nothing?
They will, if speed is the only thing you count. That is why the definition of a substantive reply should be written down, and why someone should read a sample of conversations regularly.
Where do I start without a CRM?
With mystery shopping and a manual spreadsheet for one week. That is enough to see the gaps. If you want help bringing channels together and setting up measurement, get in touch — the first conversation is about understanding your data and scope.