How Fast Do You Have to Respond to a New Lead?

What is a good lead response time?

Five minutes is the working target for lead response time, and an hour is the point where the odds change sharply.

The best-sourced number comes from a 2007 lead response management study by James Oldroyd, run with MIT's Sloan School of Management and InsideSales, which analysed three years of data across six companies, more than 15,000 leads and over 100,000 call attempts. It found that the odds of making contact with a web lead dropped by about 100 times when the call went out at thirty minutes rather than five, and the odds of qualifying it dropped by about 21 times. That comparison is worth reading twice, because it is five minutes against thirty, not against a day.

The larger audit came four years later. Oldroyd, Kristina McElheran and David Elkington published The Short Life of Online Sales Leads in Harvard Business Review in March 2011, based on test web inquiries sent to 2,241 US companies. Of those, 37 percent responded within an hour, 16 percent within a day, 24 percent took longer than a day, and 23 percent never responded at all. Among the companies that responded within thirty days, the average took 42 hours. Firms that made contact inside the first hour were nearly seven times as likely to qualify the lead as those that made contact an hour later, and more than sixty times as likely as companies that waited 24 hours or longer.

Be careful with the rest of the numbers in this category. Plenty of the statistics circulating about speed to lead, including the widely quoted claim that 78 percent of buyers purchase from the first company that responds, trace back to aggregator blogs citing each other rather than to a published study. The two above have named authors, a sample, and a method, which is why they are the two on this page.

Both studies are old, and that is not the weakness it looks like. What has changed since 2011 is the number of places a request can arrive from, so the coverage problem is larger in 2026 than it was when the audit ran, not smaller.

What actually happens when you take an hour to call back?

What happens in that hour is that the lead keeps shopping, and by the time you call, somebody else has already booked the job.

The mechanism is not mysterious. Someone with a leaking roof or a dead air conditioner fills in three forms in ten minutes, because they are not shopping, they are solving a problem today. Whoever gets there first is talking to a person who still has a problem. Whoever gets there in four hours is talking to a person who has an appointment booked and is now being polite.

That is what the HBR study is measuring when it reports a sevenfold difference in qualification odds for the firms that got there inside the hour. The lead did not get colder in some vague way. It got taken.

For a home services business the pattern has a second layer, which is that the busiest hours to answer are the same hours your crews are on jobs. Calls arrive while the phone person is dispatching, forms arrive at nine at night, and the weekend has nobody on it at all. The response time on your slowest channel is the one that decides what happens to the marketing you already paid for.

The number worth putting on this in your own business is simple to get. Take last month's leads, find the timestamp each one arrived and the timestamp of the first real reply, and look at the median. Owners are usually wrong about this figure by a factor of five, because they are remembering the leads they answered fast.

Why do most businesses take hours to respond?

Most businesses take hours to respond because responding fast is a staffing problem, and staffing it the obvious way costs more than the leads are worth.

To answer every inquiry in five minutes with people, you need somebody available every minute you accept leads. That means evenings, because form fills do not stop at five, plus weekends, plus cover for lunch and holidays and the twenty minutes your office manager is on the phone with a supplier. Very few operations under fifty people can justify that seat, so they do the reasonable thing and answer in batches, and the batch is where the leads die.

Here is how the realistic options compare on the two things that decide the outcome, coverage and whether the reply completes a booking.

Approach	Covers nights and weekends	Completes the booking	Where it falls down
Office staff answering between jobs	No	Yes	Batching. Replies go out when somebody gets a gap
An extra seat hired to cover the phone	Only the hours you pay for	Yes	Cost per lead, and the seat is idle in slow weeks
Answering service	Usually	No. Takes a message	Someone still has to call back, qualify and book
Autoresponder or basic chatbot	Yes	No	The customer has a receipt and keeps shopping
Automated lead response system	Yes	Yes	Needs real integrations and an escalation path

The second reason is that the reply has to do something. An instant autoresponder saying "thanks, we will be in touch" is not a response. The person is still shopping and now they have a receipt. What changes the outcome is a real exchange: what is the problem, where is the property, is it inside the service area, when can somebody come, and then an actual appointment on an actual calendar. That distinction between a system that responds and one that acts is the whole game here.

Can AI respond to leads without embarrassing you?

AI can respond to leads without embarrassing you, if it is built to act rather than to chat and if it is fenced in properly.

The failure mode everyone has seen is a bot that invents an answer. A system handling real inquiries should not be free to say whatever sounds right. In the systems we build, the model decides what to say and which action to take, and the code decides what is allowed: it can only offer appointment times that are genuinely open on the calendar, it checks the job address against the real service area before it offers anything, and it is not permitted to state a price that did not come from the business's own numbers.

Then it needs an exit. When the job is out of area, commercial, unusual, or when the customer asks for a person, the right behavior is to hand over with the full conversation attached rather than to keep trying. A system that never escalates is a system that will eventually promise something you cannot deliver.

What that produces is a reply in seconds, a real conversation over text, a qualified job, a booked slot on the calendar your team already uses, and a notification to whoever is running the schedule. Nobody in the office has to touch it unless the system asks them to.

If you want to see the standard of delivery behind that, the Universidad Maimonides case study is the closest public example: a narrow, high-volume, repetitive job, built in two weeks, running ever since on our infrastructure and returning about fifteen hours a week.

How do I know if it is actually working?

You know it is working by measuring three things, and by measuring them before you build anything so you have a baseline.

First, median time to first meaningful reply, split by channel and by hour of day. The median matters more than the average, because one lead answered in three days will drag an average around and hide the pattern. Splitting by hour is what tells you whether the problem is capacity during the day or coverage at night.

Second, the share of inquiries that receive a real reply at all. In the HBR study, 23 percent of the 2,241 companies never responded to a test inquiry. Most owners would not guess their own number is above zero, and it usually is, because leads arrive in more places than anyone tracks: the form, the ad platform, the missed call, the voicemail nobody plays.

Third, booked appointments per hundred leads. That is the number the first two exist to move, and it is the one that tells you whether faster replies turned into work on the board rather than into faster rejections.

Our own answer to keeping this honest is a monthly report showing what the system handled and what it produced, so you are never taking our word for whether it is still earning its place. Whoever builds yours, insist on the equivalent, and insist that the numbers come out of the system rather than out of a slide.

Frequently asked questions

How much does it cost to automate lead response?

It depends on how many channels the leads arrive on, whether the calendar and CRM you use have real integrations, and how much of the conversation has to complete without a person. Wiring one web form to text replies and one calendar is a small build. Covering forms, ad platforms, missed calls and voicemail across several locations is a larger one. The figure to establish first is what you currently spend to generate the leads that go unanswered, because that is the cost the system is offsetting.

How does an AI lead response system actually work?

A lead arrives through a form, an ad platform or a missed call, and a webhook hands it to the system within seconds. The system texts an acknowledgement, then works the conversation: what the job is, how urgent it is, where the property is. It checks the address against the service area, offers only times that are genuinely open on the calendar, books the appointment, and notifies the owner and the assigned tech. Anything out of area, commercial, or unusual gets handed to a person with the transcript attached.

Is this better than hiring an answering service?

They solve different halves. An answering service picks up the phone and takes a message, which still leaves someone to call back, qualify, and book. A lead response system completes the booking and writes it into the calendar your team already opens, and it covers text and form inquiries as well as calls. Plenty of businesses run both, with the service handling voice and the system handling everything that arrives in writing.

Can my lead process be automated if my jobs are all different?

Usually yes, because the part being automated is the first ten minutes, not the estimate. Qualifying a job, confirming the address is in your area, and putting a visit on the calendar follows rules you already apply, even when the work itself varies enormously. The pricing, the diagnosis and the judgment stay with your people.

Will customers know they are talking to a system?

Assume they will work it out, and build accordingly. The systems that hold up are useful rather than convincing: quick, accurate about what you do and where you travel, and straight about handing over to a person when the question is beyond them. Nobody minds being booked in two minutes at nine at night.