The short guide to this is buy leads or build your own system. This post is the longer argument: what each model really costs per customer, the shape of that cost over twelve months, and the rule for running both without fooling yourself.
What's the difference between buying leads and generating your own?
A bought lead is a contact someone else generated and sold to you, either as a shared lead (sold to more than one business) or an exclusive lead (sold to you alone). A generated lead comes from attention you earned or paid for yourself: your ads, your content, your Google Business Profile, your referrals, pointed at pages and follow-up you control. The test is simple. When you stop paying, what's left?
| Bought leads | Your own system | |
|---|---|---|
| Who else gets the lead | Often other businesses, if shared | Only you |
| What you control | Almost nothing before the lead lands | Message, audience, offer, follow-up |
| What you keep when spend stops | A list of past contacts at best | Retargeting audiences, content and proof, lead-to-sale data |
| Speed to start | Days | Weeks |
| Cost per customer over time | Flat, set by the marketplace | Starts higher, can fall as it compounds |
| Main risk | You pay whether or not you win | You build it and it doesn't beat the bought source |
That last row matters. Building is not automatically better. It's better if it gets your cost per customer below what you'd pay a marketplace, and not otherwise. Everything below is about finding out which.
What does each one really cost per customer?
Whichever way you get a lead, the cost that counts is the same division I covered in what lead generation costs in 2026: cost per customer = price per lead ÷ the share of leads that become customers. Bought leads change one number in it, and it's the denominator.
A shared lead isn't yours alone. If N businesses receive the same enquiry and the customer hires one, the average business wins one in N before anyone's speed or reputation is counted. At N = 3 that's 33%. At N = 5 it's 20%. Real win rates run lower, because some customers hire nobody at all. Your price per lead is the same whether you win or lose, so cost per customer is the price divided by that win rate. A $60 lead won one time in five costs $300 per customer in lead fees alone.
Speed is what moves you from the average to the front. InsideSales' study of 5.7 million leads found conversion was 8 times higher when the first call came within five minutes, and under 1% of first attempts happen that fast. When you're one of several buyers, the five-minute rule decides who gets the share (the evidence is here). It's why the same bought lead can be a good deal for the business that answers in two minutes and a bad one for the business that answers tomorrow.
Your own leads have a different problem. They cost more to start, because nobody has learned yet which message, audience and offer work for your business. You pay for that learning. What you get in exchange is a lead only you have, arriving with context you wrote.
What does the cost look like over twelve months?
Here's the part most comparisons skip. Bought leads and built systems don't just differ in price. They differ in shape. A bought source tends to be flat: the marketplace sets the price and you convert at your usual rate. A built system is a curve. It starts behind, and what happens next is the whole question.
An illustrative model. Same $4,000 a month all-in on both, over twelve months. These are assumptions, not results: they're there so you can swap in your own.
- Bought: $80 a lead, 6% close, so $1,333 per customer, flat. That's 3.0 customers a month.
- Built: months 1 to 3 at $1,800 per customer (you're paying to learn), months 4 to 6 at $1,300, months 7 to 12 at $900 as audiences, proof and follow-up start working together.
| Cumulative customers | Month 3 | Month 6 | Month 8 | Month 12 |
|---|---|---|---|---|
| Bought, $1,333 flat | 9.0 | 18.0 | 24.0 | 36.0 |
| Built, $1,800 → $1,300 → $900 | 6.7 | 15.9 | 24.8 | 42.6 |
On those assumptions the built system is behind for seven months, overtakes in month 8, and finishes the year about 18% ahead: 42.6 customers against 36. And it exits the year producing 4.4 customers a month against 3.0, on the same spend.
Two things to take from it. First, the cost of building is real and comes first. A business that judges both at month three will conclude that building lost, and it'll be right about month three. Second, the whole result rests on one assumption: that cost per customer falls. If your built system plateaus at $1,500, it never catches the bought source, and you should buy. That's the test to hold it to, and it's why the next section matters more than the model.
What has to be true for building to win?
The curve only bends down if three things are true. These aren't aspirations; they're the conditions under which a lead you generate costs less per customer than one you buy.
- Follow-up is fast and owned. Someone answers inside minutes, and it's in a person's job description. Without it a built lead and a bought lead go cold at the same speed, and you've paid more to build.
- Every lead is tracked to a sale. You can say which source produced which customer. If leads and revenue sit in different tools, you can't see the curve at all, and you can't tell a good build from a bad one. This is the most common reason founders can't answer the buy-or-build question: they have no cost per customer for either. Leads-but-no-sales is the diagnosis if that's you.
- The offer and proof can persuade a stranger. A built system sends cold people to pages and videos you control. If your proof doesn't hold up there, cost per customer won't fall, because the thing compounding is trust.
When those three are true, the compounding is mechanical. Retargeting audiences are made of people who already watched or visited, so you stop paying to restart each conversation from cold. A library of proof means each new buyer arrives half-persuaded. Lead-to-sale data tells you which message to repeat. None of it exists in a marketplace's account. It exists in yours, and it stays when the spend stops.
Exclusive or shared: what's actually being sold?
Treat every claim about lead quality as a marketing claim until you've measured it yourself. In January 2023 the FTC ordered HomeAdvisor to pay up to $7.2 million. Its complaint alleged that since at least mid-2014 the company had made false, misleading or unsubstantiated claims about the quality and source of its leads, including overstating how often leads led to jobs. The order bars it from claiming its leads are from people "ready to hire" a provider. That isn't a point about one company. It's a point about the model: when the seller of the lead also describes its quality, the buyer's only protection is their own data.
Exclusive leads are a real improvement on shared ones, because you're no longer racing N−1 other businesses. They're also priced for it. The arithmetic doesn't change: an exclusive lead at twice the price is the better buy only if it closes at more than twice the rate. Use the same division, and run it per source, not per vendor claim.
When buying leads is the right call
There's a legitimate case for the bottom of the ranking, and I laid out where it sits in how to get leads for your business. Buying makes sense when:
- You need to learn whether a market buys at all. A bought source gives you an answer in weeks, before you've built anything. That information is worth paying a premium for.
- You have capacity to fill this month. A new crew, a gap in the diary. A built system won't pay back inside the quarter. A bought source can.
- Your follow-up is already fast. You answer in minutes, you qualify properly, and you can already track to a sale. A bought source rewards exactly that.
- You're early. Under roughly $3,000 to $5,000 a month in ad spend, there isn't enough volume for a built system to learn from, which is the floor in how much to spend before hiring an agency.
The mistake isn't buying leads. It's building the business on them and wondering why cost per customer never improves. A bought source is a bridge. It should have a budget cap and a review date.
When building is the wrong move
- You can't track a lead to a sale yet. Fix that first. It's cheaper than any build, and without it you won't know whether the build worked.
- Nobody can follow up within minutes. More leads of any kind will go cold in the same inbox.
- Your offer isn't proven. A built system amplifies an offer that already sells. It can't create one.
- Referrals fill the calendar. Then you don't have a lead problem, and neither source is your priority.
How do you run both without fooling yourself?
Most established businesses end up here, and it works if you run it as a comparison rather than a belief. A rule I'd use:
- Cap the bought source. Set a fixed monthly budget you'd be comfortable never increasing.
- Build alongside it, pointed at the same offer. Same offer, same follow-up, same tracking, so the comparison is fair.
- Review at 90 days and again at 180. Cost per customer per source, from lead to signed. Not cost per lead, and not what either vendor says.
- Fund the lower number. If the built system is above the bought source at month six and the curve isn't bending, say so and shift the budget. If it's below, move budget across and keep the bought source as a capped test channel.
The Compare Funerals account shows how quickly the numbers on an owned account can move. Compare Funerals is a UK funeral-plan provider serving British expatriates in Spain, so these figures are in pounds. In our first 30 days its cost per qualified lead fell from £521.07 to £147.65, a 71.7% drop, while qualified opportunities rose from 24 to 60 a month on 29.5% less ad spend. That's a fix to an existing account, not a build from nothing, and it's an acquisition-efficiency case study, not a revenue one. But it's the same mechanism the model above relies on: the cost per qualified opportunity is something you control once the system is yours. For attention that turns into enquiries, an AI-led campaign for Ironclad Finance produced more than 5 million views and 74 inbound enquiries in seven weeks.
Owning the loop is what I build for a living. The lesson from $60M+ in tracked revenue across 96+ clients is less about which source wins and more about whether you can see the number for each. We don't publish a price: it depends on your ad spend, what a customer is worth to you and your capacity. It's a one-time installation plus monthly management, and your ad budget stays yours. If you want the division done on your own sources, Book A Free Call.
Frequently asked questions
Should a service business buy leads or build its own lead generation system?
Both, with the bought source capped, until the numbers say otherwise. Buy to fill capacity or test a market quickly. Build to make cost per customer fall over time. Judge each on cost per customer, lead to signed sale, over 6 to 12 months. A built system usually costs more per customer for the first few months before it overtakes, so the decision has to be made on that curve, not on month one.
Are bought leads worth it?
Sometimes. A bought lead is worth it if its cost per signed customer sits below your ceiling for what a customer is worth, and that needs tracking from every source to a sale. Shared leads tend to close lower, because several businesses compete for the same enquiry and speed decides who wins. If you can answer within minutes and track the outcome, the answer is more often yes.
How long does it take for your own lead generation to beat bought leads?
It depends on spend, offer and follow-up, so there's no honest single number. In the illustrative model above it takes about eight months. In a business with tracking and fast follow-up already in place it can be faster. In one with neither it may never happen, which is why the three conditions come first.
What do I own if I build my own system?
Retargeting audiences made of people who engaged, a library of proof and content, and lead-to-sale data showing which message produces customers. A marketplace owns all of that on its side, and you keep only the contacts it chose to send.
Sources
- InsideSales, Lead Response Management 2021 (5.7 million leads; conversion 8 times higher within five minutes; under 1% of first attempts that fast).
- U.S. Federal Trade Commission, FTC Order Requires HomeAdvisor to Pay $7.2 Million, Stop Deceptively Marketing Its Leads to Home Improvement Professionals (23 January 2023).
- The twelve-month comparison and the shared-lead win-rate figures are illustrative arithmetic on stated assumptions, not published benchmarks or client results.
