The only audited lead conversion rate in home services is legally sealed

A federal regulator subpoenaed a lead seller's own conversion data, found the published rates could not be substantiated, and redacted the real numbers as confidential. Meanwhile the market leader's audited filings put the average lead at $25 to $33, not $50 to $150.

If you buy leads for a roofing, HVAC or home-improvement business, you have been quoted a price and a conversion rate. The price is usually $50 to $150 a lead. The conversion rate is usually attached to a comparison: exclusive leads convert three to five times better than shared ones, so pay more.

We went looking for where either number was measured. What we found instead is the most useful fact in this entire category, and almost nobody in it mentions it.

That is the state of the evidence. The only lead-to-job conversion rate in home services that has ever been audited by anyone other than the party selling the leads exists, was produced under subpoena, and is confidential. Every conversion figure you have been quoted sits on top of a number that is legally sealed.

What the leads actually were

The complaint's Count I is about whether the leads were what they were sold as. Paragraph 32 is the mechanism, and it is worth reading slowly:

HomeAdvisor sells information to service providers as leads even when internet users affirmatively indicate that they do not intend to hire a service provider soon.

FTC administrative complaint, In the Matter of HomeAdvisor, Inc., 11 March 2022, paragraph 32

The form asked the homeowner to pick “Ready to Hire” or “Planning & Budgeting”, and a timeframe. HomeAdvisor did not dispute that those questions existed: the Commission's September 2022 order listing facts without substantial controversy records them as established, each citing the company's own answer to the complaint.

Two proportions survived redaction, and both are worth having. Roughly 5 to 10% of leads sold were generated over the telephone rather than from a web form, per a corporate witness's sworn testimony. And fewer than 5% of leads sold were Instant Booking leads, the product where a homeowner picked a named contractor and booked an appointment. The closest thing in the catalogue to a genuinely exclusive, high-intent lead was under one lead in twenty.

What a lead actually sells for, from audited accounts

The price is easier, because the market leader is a public company and has to file. Angi Inc. reports US lead revenue and a count of leads in the same table of the same 10-K. Dividing one by the other gives an average revenue per lead that is audited on both sides.

Fiscal yearUS lead revenueLeadsRevenue per leadContractors per homeowner request
FY2022$954.7m28.9m$32.990.98
FY2023$781.1m27.1m$28.811.17
FY2024$606.6m24.4m$24.881.42
FY2025$587.1m20.2m$29.071.30
Angi Inc. annual reports on Form 10-K, US segment. Angi renamed these metrics between the FY2024 and FY2025 filings: FY2024's “Consumer connection revenue” is FY2025's “U.S. Lead revenue” at the identical figure, and “Monetized Transactions” became “Leads” at the identical count, so the series is continuous. The last column is Leads divided by Service Requests, which Angi's own definitions make meaningful: “a single Service Request can result in multiple Leads.” FY2022 is not comparable on that column, because the older definitions counted different things.
  • $50The low end of the $50-$150 range quoted across contractor marketing. No publisher of it states a sample.
  • $32.99Angi FY2022 10-K: US lead revenue divided by leads
  • $28.81Angi FY2023 10-K, same construction
  • $24.88Angi FY2024 10-K, same construction
  • $29.07Angi FY2025 10-K, same construction

Quoted by parties selling leads or lead softwareAudited, filed with the SEC

Cost of a home-services lead. The first bar is the BOTTOM of the range in general circulation, plotted at its most generous end, and every audited year still comes in below it. Two caveats printed rather than buried: this is a blended average across every category Angi sells, from house cleaning to roofing, and roofing sits at the expensive end, so a roofing lead is above this line. And it is revenue per lead delivered rather than price per lead quoted, so it nets out the credits Angi actually gave back.

The exclusivity multiplier, measured

The last column of that table is the only audited public measure of lead sharing in the industry, and it is there by accident. Because Angi defines a Service Request as a homeowner's request and a Lead as a connection resulting from one, and states plainly that a single Service Request can produce multiple Leads, dividing the second by the first tells you how many contractors each homeowner was sold to.

At its peak, in FY2024, that figure was 1.42. The average homeowner request went to roughly one and a half contractors, not five. Set that beside the standard pitch that exclusive leads convert three to five times better than shared leads: the shared product being described is, on the market leader's own audited numbers, mostly shared with one other person.

The sentence Angi files every year

The clearest description of what a lead is comes from the seller, in the section of the 10-K where the SEC requires it to describe its own business. Angi has described its revenue, for four consecutive audited years, as:

fees paid by Pros for consumer matches (regardless of whether the Pro ultimately provides the requested service)

Angi Inc., annual report on Form 10-K, FY2025, Item 1

The parenthetical is the whole thing. You are not buying an outcome and the filing says so. The same document's risk factors add that “the quality and convertibility of traffic and leads generated through third-party arrangements are dependent on many factors, most of which are outside our control.” The FTC noticed that sentence too, and quoted it back at the company in paragraph 46 of the complaint.

What the order changed, and what it did not

The case settled. The final consent order, approved 21 April 2023 and running for twenty years, forbids HomeAdvisor from describing leads with phrases including “Ready to hire”, “Project-ready”, “Not window shopping” and “Not just price-shopping”. It also sets a substantiation rule that is the single most useful sentence for anybody buying leads from anybody:

unless the representation is non-misleading and Respondent (1) has a reasonable basis for the claim at the time it is made; (2) has in its possession data or written materials to substantiate the claim

FTC Decision and Order, 21 April 2023, provision I.B, on claims about the rate at which leads convert into jobs

That order binds one company. Nothing in it stops any other lead seller, marketplace, agency or software vendor from quoting you a conversion rate it cannot evidence. The rule that a claim needs data behind it before it is made is not a general obligation in this market. It is a penalty applied to one party that was caught.

The redress is the part worth remembering. The FTC sent 110,372 checks to home service providers. The order caps the leads fund payments at up to $30 per membership. Angi's own filings put what a contractor spends with it at roughly $5,750 a year, stable across three years. A contractor found to have been misled about what they were buying was made whole to the extent of about half a per cent of one year's spend.

The two defects Google stopped refunding

The FTC order binds one company. But the two specific defects it sued over turn out to be, at the largest advertising platform selling leads to these same contractors, explicitly non-refundable.

Google Local Services Ads charge per lead, and Google publishes what it will and will not credit back. On its own support page, under the heading About Automated Local Services Ads lead credits:

Google no longer supports credits for “job type not serviced” and “geo not serviced” leads.

Google Local Services Help, About Automated Local Services Ads lead credits

Now read those two categories against paragraph 62 of the FTC complaint, which alleges that leads sold to contractors “did not concern projects that match the types of services that service providers have expressed they perform” or “did not concern projects that match the geographic areas that service providers have expressed they serve”. Those are the same two failures, named almost the same way. One platform was sued for selling them. The other has stopped giving the money back for them.

Google's published list of what will not be credited also includes “A customer was researching potential projects or prices related to a service you offer”, which is the same population the FTC described as internet users who affirmatively indicated they did not intend to hire soon. And on price, Google's own documentation says only that “Lead prices may vary depending on your location, the job type, the type of lead, or your bidding mode”. It publishes no rate card at all. Every Local Services cost-per-lead figure you have read is somebody's estimate of a number the seller does not disclose.

What this means if you buy leads

We build capture and response systems for roofers and home-services contractors, so the disclosure is that we work on the thing that happens after a lead arrives, and we have an interest in you thinking that part matters. Which is exactly why there is no cost-per-lead figure and no conversion rate anywhere else on this site.

The practical version:

  1. Ask for the substantiation, in writing, before you buy. The FTC order spells out the standard: a reasonable basis, and data in hand, at the time the claim is made. A seller who cannot produce it when asked is telling you something.
  2. Price the lead against your own close rate, not theirs. Yours is knowable from your own records this quarter. Theirs is, in the one case anybody ever checked, sealed.
  3. Count how many contractors the enquiry went to. Ask the homeowner on the first call. It takes one question and it is the only version of that number you will ever be able to verify.
  4. Separate the lead cost from the membership cost. The published price per lead excludes the annual fee, and the fee is not small relative to the leads.
  5. Measure what you do with them. The one variable in this that is fully inside your control and fully measurable is how fast and how consistently a lead gets worked once it lands. Nobody has to substantiate a claim to you about that, because it is your own data.

None of this says lead marketplaces are worthless. A shared lead at $29 that you work properly can be excellent value, and for a contractor with no other channel it often is. It says that the two numbers you are budgeting against, the price and the conversion rate, are respectively higher than the audited figure and unknowable, and that the one organisation that ever checked found the seller could not stand behind what it was claiming.

Sources

Publisher, year, sample, and what the publisher sells. The same four things every cited figure on this site carries.

In the Matter of HomeAdvisor, Inc. (administrative complaint)

Publisher, year and sample
Federal Trade Commission, Docket No. 9407, issued 11 March 2022, public redacted version, 17 pages. Count II is titled “Misrepresentations About the Rates at which Leads Convert into Jobs”; paragraph 6 alleges leads were represented as converting “at rates well above what HomeAdvisor can substantiate”. Paragraphs 49 to 51 give the company's own win-rate ranges since July 2014 with every numeric value redacted, as does the Commission's Opinion Following Order Denying Summary Decision of 28 September 2022. Paragraph 32 alleges leads were sold even where users indicated they did not intend to hire soon; the underlying form questions were recorded as established facts in the Commission's 9 September 2022 order, each citing HomeAdvisor's own answer.
What the publisher sells
The FTC is the federal regulator and sells nothing. This is the only instance we found in which any party other than a lead seller examined a lead seller's own conversion data and published a finding about it.

FTC Decision and Order (final consent), HomeAdvisor

Publisher, year and sample
Approved 21 April 2023, twenty-year term. Provision I.A prohibits describing leads with phrases including “Ready to hire”, “Project-ready”, “Not window shopping” and “Not just price-shopping”. Provision I.B prohibits any representation about the rate at which leads convert into jobs unless the respondent has a reasonable basis and data in its possession to substantiate it at the time the claim is made. Provision III caps leads-fund payments at up to $30 per membership. The FTC's press release of 28 November 2023 states it sent 110,372 checks to eligible home service providers.
What the publisher sells
Primary regulatory document. Note the limit of its reach: it binds one company, and creates no general obligation on any other lead seller to evidence a conversion claim.

About Automated Local Services Ads lead credits

Publisher, year and sample
Google Local Services Help, read 29 August 2026. States that Google “no longer supports credits for ‘job type not serviced’ and ‘geo not serviced’ leads”, which are the same two defects pleaded at paragraph 62 of the FTC complaint against HomeAdvisor. Google's companion page on how leads work lists further non-creditable cases including a customer who “was researching potential projects or prices related to a service you offer”, and says of price only that it “may vary depending on your location, the job type, the type of lead, or your bidding mode”. No rate card is published.
What the publisher sells
Google sells the leads. What makes the page usable is that it is the seller's own binding statement of what it will refund, published as product documentation rather than as marketing.

Angi Inc. annual reports on Form 10-K, FY2022 to FY2025

Publisher, year and sample
Audited filings. US lead revenue and lead counts appear in the same table: $954.7m over 28.9m leads (FY2022), $781.1m over 27.1m (FY2023), $606.6m over 24.4m (FY2024), $587.1m over 20.2m (FY2025), giving $32.99, $28.81, $24.88 and $29.07 per lead. Leads divided by Service Requests gives 1.17, 1.42 and 1.30 contractors per homeowner request for FY2023 to FY2025. Angi renamed the metrics between the FY2024 and FY2025 filings at identical figures, so the series is continuous. Item 1 describes revenue as “fees paid by Pros for consumer matches (regardless of whether the Pro ultimately provides the requested service)”. Ads and Leads revenue divided by transacting professionals gives roughly $5,750 a year per contractor across FY2022 to FY2024.
What the publisher sells
Angi sells the leads, so it is not a disinterested party. What makes these figures usable anyway is that they are audited, filed under securities law, and cut against the seller's own marketing: a lower price per lead and a smaller exclusivity gap than the category advertises. The blended average spans every category Angi sells, and roofing sits above it.

We would ratheruse your numbers.

Every article here ends the same way, because the honest answer is always the same: the benchmark is somebody else’s and your own figures are the ones that decide anything. On the call we work them out with you.

Fourteen questions, about seven minutes. No price, no purchase, and nobody calls you unless you ask them to.

Application1 / 14

Next question: where we send what we prepare.