How to Automate Lead Follow-Up So No Lead Goes Cold

You automate lead follow-up by connecting every place a lead can arrive to one system that replies instantly, tells a person, and keeps following up on a schedule until the lead answers, books or opts out. The automation handles speed and memory. A person handles the conversation. Done well, a missed call at 4:55 p.m. gets a text by 4:56 p.m. and a call back before 5:05 p.m., whether or not anyone remembered.
Here is the short version. Leads go cold because nobody answered in the first hour, not because the owner did not care. Most businesses already pay to generate leads through Google Ads, Maps and referrals, then lose a share of them to a voicemail box or an inbox nobody checked until the evening.
I set up these systems for local service businesses, mostly trades, clinics and home services, and the pattern is always the same. The ad account gets the attention. The follow-up gets whatever is left of the owner’s day. This guide is the follow-up system I actually build: where to catch each lead, what happens in the first five minutes, the sequence by job type, when to stop, the consent rules, the tools, and the numbers that prove it is working.
Why do leads go cold so fast?
Leads go cold because interest decays in minutes and most businesses respond in hours. The best-known audit, published in Harvard Business Review in 2011, tested 2,241 US companies with a web lead. Only 37% replied within an hour, and 23% never replied at all. The buyer does not wait. They call the next business on the list.
The authors, Oldroyd, McElheran and Elkington, put it plainly: “Although 37% responded to the lead within an hour, and 16% responded within one to 24 hours, 24% took more than 24 hours and 23% of the companies never responded at all.” Among companies that did respond within 30 days, the average response time was 42 hours.

What the research actually measured
The same HBR piece analysed 1.25 million leads and found that firms contacting a lead within an hour were nearly seven times as likely to qualify it as firms that waited even one hour longer, and more than 60 times as likely as firms that waited 24 hours or more. The separate Lead Response Management study, led by Dr James Oldroyd with InsideSales data from more than 15,000 leads, found a 21-fold drop in the odds of qualifying a lead when response time stretched from 5 to 30 minutes.
Two honest caveats. Both studies are more than a decade old, and both measured mostly sales-led companies rather than plumbers or dentists. I still use them because nothing newer and better has replaced them, and because the direction matches what I see in call logs: the enquiries answered while the customer is still holding the phone book far more often than the ones answered after lunch.
The local service version of the problem
Home services are not better at this. A 2018 mystery shopping test by the agency Valve+Meter enquired with 466 home-service companies. After five days, only 60% had responded, and 95% had not replied within five minutes. That is a secondary, older dataset, so treat it as a signal rather than a benchmark.
In local trades the cause is rarely laziness. The owner is on a roof, under a sink or with a patient. Calls go to voicemail, web forms land in a shared inbox, and Google Ads lead forms sit in a spreadsheet nobody opens. The failure is the system, not the person, which is exactly why automation fixes it.
What does automated lead follow-up actually mean?
Automated lead follow-up is software that responds to a new enquiry the instant it arrives, notifies a person, and sends a scheduled series of texts, emails and call reminders until the lead replies or the sequence ends. It is one part of lead management, sitting between lead generation and the sales conversation.
It borrows from marketing automation and drip marketing, but for a local service business it is much simpler than either. You do not need lead scoring models or a 14-email nurture. You need three things to happen every time, for every lead, at 2 p.m. on a Tuesday or 9 p.m. on a Sunday:
- An instant acknowledgement, so the lead knows a real business got the message.
- A human notified, with enough detail to call back without hunting.
- A memory that does not forget, so the third and fourth touches happen even in a busy week.
| Approach | First reply | Follow-up memory | Where it breaks |
|---|---|---|---|
| Manual | Whenever someone checks | Sticky notes and memory | Busy days, evenings, weekends |
| Automated | Under 60 seconds | Scheduled, never skipped | Robotic copy, no stop rules |
| Automated + human | Under 60 seconds, call in minutes | Scheduled, stops on reply | Only if nobody owns the alerts |
| Fully AI | Instant, conversational | Scheduled | Quotes, complaints, judgement calls |
What to automate and what to keep human
My rule is simple: automate anything that depends on speed or memory, and keep anything that depends on judgement. The automated half:
- The first text and email to every new lead.
- Alerts to the owner, the front desk or the on-call tech.
- Reminders for unanswered leads and unaccepted quotes.
- Appointment confirmations and the day-before reminder.
- Moving the lead to the right stage in the CRM.
The human half is the actual conversation: pricing, scope, whether a site visit is worth it, and anything that starts with a complaint. A chatbot can collect a postcode and a photo. It should not be the one telling a customer why their quote went up.
Where do your leads come in, and how do you catch every one?
A typical local service business has five to eight lead sources, and automation only works on the ones connected to it. List every door a lead can walk through, connect each one to a single system of record, and test each with a fake lead. In audits, the leak I find most often is not a slow reply. It is a source nobody connected at all.
| Lead source | How to catch it | Common leak |
|---|---|---|
| Phone calls | Call tracking or a business line with missed-call detection | Voicemail nobody checks until evening |
| Website forms | Form posts straight into the CRM | Form emails a shared inbox |
| Google Ads lead forms | Native integration or the lead form webhook | Leads left in the Ads interface or a CSV |
| Local Services Ads | Answer calls live, reply to message and booking leads fast | Missed calls that hurt responsiveness |
| Meta lead ads | Native CRM sync or an automation tool | Downloaded weekly, by which time they are cold |
| Directories and marketplaces | Email parsing or the platform’s integration | App notifications on one person’s phone |
Missed-call text-back: the first automation to build
If you build one thing this week, build this. When a call to your business number goes unanswered, the system texts the caller within seconds. Back in 2016, the directory network 411 Locals reported that only 37.8% of calls to the small businesses it tested were answered by a person. Even if your rate is double that, a meaningful share of callers are hearing a voicemail greeting and dialling the next result.
A missed-call text turns that dead end into a conversation. Many callers prefer to reply by text anyway, especially when they are at work themselves. One rule: the text must come from the same number they called, or it reads like spam. To put a rough figure on what unanswered calls cost you, try my missed-call calculator.
Webhooks: the plumbing behind instant follow-up
A webhook is a message one system sends another the moment something happens. It is how a lead from an ad platform reaches your CRM in seconds instead of hours. Google documents a lead form webhook for lead form assets in Google Ads, so leads can post straight into your system.
If your CRM has no native connector, a tool like Zapier can catch the webhook and pass it on. It adds a link to the chain, so test it monthly. A broken connection fails silently, and you only find out when the lead count drops.
One channel that no longer exists
Older guides still tell you to switch on chat in your Google Business Profile. Google retired Business Profile chat in July 2024, so that door is closed. Calls, the website and your booking link carry the load from the profile now, which makes the missed-call text even more important.
What should happen in the first five minutes?
In the first five minutes, the lead should get an automatic text, a person should get an alert with a one-tap way to call, and the lead should get a short email with what happens next. The text buys time. The alert gets a human on the phone. Aim for a person calling back inside five minutes during business hours, since that is where the steepest drop in the research sits: the Lead Response Management study measured a 21-fold fall in qualification odds between minute 5 and minute 30.

In order, the first five minutes look like this:
- Second 0 to 60: the lead gets a text from your business number, using their name and the service they asked about.
- Same minute: the owner, front desk or on-call tech gets an alert with the lead’s name, number, source, job type and a tap-to-call link.
- Minute 1 to 5: a person calls. If they cannot, they reply to the text with a time they will call.
- Minute 5: an email lands with what happens next, a price range if you publish one, a few job photos and a booking link.
- Minute 15: if nobody has claimed the lead, the alert escalates to a second person.
What the first text should say
Short, specific and signed. Name the business, name the job, say when a person will call, and ask one easy question. For a plumbing enquiry, something like:
Hi Sarah, it’s Dan from Northside Plumbing. Got your message about the leaking hot water system. I’m on a job and will call you within 15 minutes. Is water still leaking right now? Reply STOP to opt out.
Notice what it does not do. It does not say “Thank you for your enquiry, a team member will be in touch.” Customers can smell that sentence. One question with an easy answer gets a reply far more often than a list of five, and the answer tells you whether this is an emergency.
The owner alert and escalation
The alert is the step most DIY setups skip, and it is the one that matters most. An automatic text with no human behind it just delays the disappointment. Send the alert by text or push notification, not email, and include everything needed to call without opening a laptop.
Then add escalation. If the lead is not marked as contacted within 15 minutes, the alert goes to a second person. In teams of two or three, this one rule does more than any message template, because it removes the assumption that someone else already called.
After-hours leads
For leads that arrive outside working hours, change the promise rather than the speed. The instant text still goes out, but it says when you will call: “first thing at 7:30 tomorrow” or “if water is coming through the ceiling, call the emergency line now”. If you offer online booking, put the link in that text. A lead who books themselves overnight is the best outcome of all.
What does a good follow-up sequence look like for a local service business?
A good sequence matches the job type, front-loads the touches into the first three days, mixes calls, texts and email, and stops the moment the lead responds. For most local service enquiries that means five to eight touches over about 30 days. Emergency jobs compress into hours, while quote-based jobs stretch over weeks. The HBR audit found the average responding company took 42 hours to reply, so even a modest sequence that starts inside a minute puts you ahead of most competitors in the Harvard Business Review data.
| Job type | Window | Touches | Goal |
|---|---|---|---|
| Emergency (burst pipe, no heating, lockout) | First 2 hours | Text, call, call, text | Get someone on site today |
| Quote request (install, renovation) | 14 days | 5 to 6 | Book the site visit |
| Quote sent, not accepted | 21 days | 3 to 4 | A yes, a no, or a reason |
| Went quiet | 30 to 90 days | 1 a month | Stay remembered, then close the loop |
Emergency jobs: compress everything
Someone with water coming through a ceiling is ringing three plumbers at once. The sequence is an instant text, a call inside five minutes, a second call at 15 minutes, and a final text at an hour offering the next available slot. After that, stop. By then they have fixed it or found someone, and a fifth message only proves you were slow.
Quote-based jobs: stretch it out
A kitchen renovation or a new air-conditioning system is a considered purchase. People collect three quotes and decide over a fortnight. A sequence I use as a starting point:
- Day 0: instant text, call within the hour, email with examples of similar jobs.
- Day 1: a second call attempt, then a text if it goes unanswered.
- Day 3: a text with one useful fact, such as lead times for the parts they need.
- Day 7: an email answering the question most people ask before booking.
- Day 14: a direct, polite text asking if they still want a visit.
Quote sent but not accepted
This is the most neglected stage, and the most profitable to fix, because the lead has already met you. Follow up on day 2 to check the quote arrived and answer questions, on day 7 with anything that changes the decision, such as availability or a finance option, and on day 21 with a close-the-loop message: “Should I close this quote off, or is it still on the table?” That question gets honest answers, and a clear no is worth more than a maybe that sits in the pipeline for months.
When should the automation stop?
Automation should stop the moment a person replies, books, opts out or is marked as lost, and it should pause when a team member takes over the conversation. Stop rules are what separate follow-up that feels attentive from follow-up that feels like a bot. Most robotic sequences I review have no stop rules at all, only a schedule. In the US, the FCC rules in 47 CFR 64.1200 also let a person revoke consent by any reasonable method, so a reply like “no thanks, found someone” has to end the texts, not only the word STOP.
- Reply received: every automated step ends and the lead goes to a person.
- Booking made: the sales sequence ends and the appointment reminders begin.
- Opt-out received: no further texts, ever, on that number, and a note in the CRM.
- Marked lost or not a fit: the lead leaves the sequence, with a reason recorded.
- Sequence complete: after the close-the-loop message, the lead moves to a quiet list.
Why most DIY sequences feel robotic
The automation is rarely the problem. The copy and the missing rules are. When I rebuild a sequence, the changes are almost always the same five, and none of them need a new tool.

Write every message as if you were texting a neighbour who asked you for a quote. Use their first name, the job they asked about and your own name. Send nothing between 9 p.m. and 8 a.m. their time. If a message would sound strange read aloud, it will read strangely on a phone.
Is it legal to send automatic follow-up texts?
Generally yes, when the lead has given consent by contacting you or opting in, every message identifies your business, every message offers an easy way to opt out, and you send at reasonable hours. The detail varies by country, and texting rules are stricter than email rules almost everywhere. This is general guidance, not legal advice, so confirm your setup with someone qualified where you operate.
United States: TCPA, quiet hours and 10DLC
The Telephone Consumer Protection Act of 1991 governs automated calls and texts. The FCC rules in 47 CFR 64.1200 bar telephone solicitations before 8 a.m. or after 9 p.m. at the recipient’s location, let people revoke consent “by using any reasonable method”, and require revocations to be honoured within ten business days. The FCC’s stricter one-to-one consent rule was vacated by the Eleventh Circuit in January 2025 before it took effect, but prior express consent is still required for automated marketing texts.
There is also a carrier rule that catches many small businesses. Texts sent from a standard US business number need an approved A2P 10DLC registration. Twilio, for example, blocks messages from 10DLC numbers that are not attached to an approved campaign. Most texting platforms walk you through registration. Do it before you launch, not after the first batch fails.
Australia: the Spam Act 2003
Under Australia’s Spam Act 2003, enforced by the ACMA, a commercial text or email needs consent, must clearly identify who sent it, and must include a working unsubscribe that is honoured within five business days. Someone who sends you an enquiry has usually given inferred consent to hear back about that enquiry. Adding them to an ongoing promotional list is a different question and needs clearer consent.
UAE and other markets
In the UAE, promotional SMS rules are set through the telecom regulator and the operators, and in practice that means opt-in consent, a registered sender ID, sending windows during the day, and a clear opt-out. Transactional replies to someone who just enquired are treated differently from promotional blasts. Because the rules sit partly in operator policy, check with your SMS provider before you send anything that is not a direct reply.
Email is the easier channel
Email is regulated too, just more loosely. In the US, the CAN-SPAM Act of 2003 requires a valid physical postal address in commercial email and a working opt-out. The FTC’s compliance guide is direct: “You must honor a recipient’s opt-out request within 10 business days.” Australia’s Spam Act covers email as well as SMS.
Which tools do you need, and which can you skip?
You need three layers: a system of record that holds every lead, a messaging layer that sends texts and emails and logs calls, and connections from each lead source into it. Many platforms bundle all three. A separate email marketing platform is optional, since most CRMs send follow-up email well enough. What you can skip, at least at first, is lead scoring, AI chat on every page and multi-branch nurture flows. Get the first five minutes right before anything clever.
| Layer | What it does | Examples |
|---|---|---|
| System of record | Holds every lead, stage and note | A CRM such as HubSpot, or field-service software such as ServiceTitan or Jobber |
| All-in-one | CRM, texting, missed-call text-back and sequences in one | GoHighLevel and similar platforms |
| Messaging layer | Sends and receives texts, routes calls | Built into the above, or Twilio underneath |
| Connections | Moves leads from each source into the CRM | Native integrations, webhooks, Zapier |
How I choose for a client
If the business already runs on field-service software and uses it daily, I build the follow-up inside it, even if its automation is less flexible. A second system that nobody opens is worse than a basic one that everybody uses. If the business has nothing but a phone and an inbox, an all-in-one platform is usually the fastest route to a working system.
Underneath every choice is customer relationship management discipline, not software. One place where every lead lives, one owner for every lead, one status that is always current. Lead scoring starts to pay off later, once you have enough volume that the team cannot call everyone first, and tools like Twilio or HubSpot only earn their place when the basics already run. If you want the bigger picture of what can run on its own, my guide to building an AI agent without code covers the next layer up.
How does fast follow-up affect your Google visibility?
For Google Local Services Ads, response speed is a stated ranking input. Google’s ad ranking documentation lists “Your responsiveness to customer inquiries and requests” among the auction factors and adds: “Missed calls may negatively affect your responsiveness.” Average response time is also listed under profile quality. Slow follow-up costs the lead and can cost the next one.
For standard Google Ads and Maps, follow-up does not change rankings directly, but it changes what the clicks are worth. Two businesses paying the same cost per lead can end up with very different cost per booked job, and that difference is mostly what happens after the lead arrives. I cover the Local Services Ads side in detail in my Local Services Ads guide.
How do you measure whether it is working?
Measure four numbers per lead source: median speed to lead, contact rate, booked rate and cost per booked job. Speed and contact rate tell you whether the system works. Booked rate and cost per booked job tell you whether it makes money. Review them weekly for the first month, then monthly. If a number gets worse, the cause is usually a broken connection or an alert nobody owns.
- Median speed to lead: time from lead arrival to the first human contact. Target under 5 minutes in business hours.
- Contact rate: the share of leads you actually spoke to or texted with. Below about two in three usually points to missing sources or slow calls.
- Booked rate by source: the share of leads that became a visit or a job, split by where they came from.
- Cost per booked job: ad spend divided by booked jobs, the number that matters more than cost per lead.
- Opt-out rate: if it climbs above a few per cent, the copy or the frequency is wrong.
Take a baseline before you switch anything on. Pull the last 30 days of leads and work out roughly how fast they were answered and how many booked. Without that baseline, you cannot tell whether the automation helped or the season changed.
Which lead follow-up stats should you stop quoting?
Several of the most repeated lead follow-up stats have weak or missing sources. The 98% SMS open rate has no primary study behind it, the “100 times more likely to contact” figure traces back to vendor marketing rather than the published research, and the research everyone cites is more than a decade old. The direction holds. The precise numbers deserve caution.
The 98% figure is the clearest case. A widely shared 2016 Gartner article only said that “various sources report” it. The UK analyst firm Mobilesquared, often credited as the origin, has since revised its own view: “Our latest data suggests 55% of all SMSes are read, however, 100% of SMSes are viewed.”
Why does this matter to an owner? Because a vendor quoting 98% opens and 21 times more conversions is selling certainty that does not exist. Texting is still the fastest way to reach a lead, and fast replies still win more jobs. You do not need inflated numbers to justify it. Your own before and after numbers will be more persuasive than any study.
How I would set this up for a local service business
Start with the leak, not the software. In order, this is the build I recommend for a single-location service business, and most of it fits into one or two weeks:
- Map every lead source and send a test lead through each one.
- Pick one system of record and connect every source to it.
- Switch on missed-call text-back from your main business number.
- Write the first text and email for each job type, in your own voice.
- Set up the alert and escalation to real people, by text.
- Build the sequences with stop rules from day one.
- Check consent and registration for your country before the first message goes out.
- Measure speed to lead and booked rate weekly for a month.
Follow-up is the cheapest growth lever most local businesses have, because the leads are already paid for. If you want this built and connected to your ads and Maps listing, it is part of how I set up local SEO and paid search for clients. Get a free audit and I will show you where your leads are leaking first.
Frequently asked questions about automating lead follow-up
The questions owners ask me most before they switch it on.
Frequently Asked Questions
How fast should you respond to a new lead?
Inside five minutes, and the first reply should be automatic so it lands inside one. The Lead Response Management study found the odds of qualifying a lead dropped 21-fold when response time stretched from 5 to 30 minutes. An instant text buys you those minutes while a person gets to the phone.
How many times should you follow up with a lead?
For a local service enquiry, five to eight touches over about 30 days is a sensible ceiling, front-loaded into the first three days. Every touch should stop the moment the lead replies, books or opts out. More messages after that point do not recover the lead. They annoy them.
What is missed-call text-back?
An automation that sends a text to anyone whose call to your business number went unanswered, usually within a few seconds. It confirms who you are, says you will call back, and asks one question about the job. For trades that spend the day on tools, it is often the single highest-value automation to set up first.
Is it legal to send automatic follow-up texts to leads?
Usually, if the lead gave consent by contacting you or ticking a clear box, you identify your business, and every message offers an easy opt-out. The US adds 8 a.m. to 9 p.m. calling windows for solicitations and A2P 10DLC registration for business texting. Australia's Spam Act 2003 needs consent, identification and a working unsubscribe. Check local rules before launch.
Should the first follow-up be a text, an email or a call?
A text first, within a minute, because it reaches a phone the lead is already holding. Then a call from a person as soon as one is free, and an email with the details they will want to read later. Email alone is too slow for urgent jobs and gets buried for quote requests.
Which tools do I need to automate lead follow-up?
Three pieces: a CRM or job-management system that holds every lead, a texting and calling layer that can send automatic messages, and something to connect the lead sources to it, such as native integrations, webhooks or Zapier. Many all-in-one platforms bundle all three. Pick one system of record and route everything into it.
Does replying slowly hurt my Google Local Services Ads ranking?
It can. Google lists your responsiveness to customer inquiries as a ranking factor for Local Services Ads and states that missed calls may negatively affect it. Average response time also sits inside profile quality. Fast, consistent follow-up is one of the few ranking inputs fully under your control there.
Can AI reply to and book my leads automatically?
It can handle the first reply, answer common questions and offer booking times, and for after-hours leads that is useful. I would keep a person on quotes, complaints and anything involving price or a site visit decision. Let AI buy time and collect details. Let a human close.
About the author

Independent AI-Powered Digital Marketing Consultant
Works with businesses worldwide·5+ years in SEO, Google Ads & AI search
I am an independent consultant focused on Local SEO, Google and Meta Ads, web design, and answer-engine and generative-engine optimisation (AEO and GEO). I run every one of these systems on my own business before I recommend it, and every audit, campaign, and report is delivered by me personally, not an account manager.
Certified:Google Ads · Meta Blueprint · Google Analytics 4
Selected results: +427% organic traffic in 30 days for a US HVAC company, and 3,770 Google Business Profile calls in a year for an Australian transport client. See the full portfolio.
