A lead's sticker price is the smallest cost involved. Divide your monthly lead spend by jobs actually booked, then add the unpaid hours spent quoting and chasing. For many electricians the honest number lands at several hundred dollars per booked job plus a workweek of unpaid admin per month — which is why the model feels worse than the invoices look.
Every electrician who buys leads knows the sticker price. Almost nobody calculates the real one. Let's do it properly, once, with math you can rerun on your own numbers.
The only number that matters: cost per booked job
Platforms price per lead — a contact, not a customer. Your business runs on booked jobs. The conversion between those two is where the money disappears.
Cost per booked job = monthly lead spend ÷ jobs booked from leads. Not per lead. Per booked job. Write this on the whiteboard.
Say leads run $60–$120 each for electrical work in a metro like ours, and you close 1 in 4 — a decent rate on shared leads, where several contractors race for the same homeowner. At $80 average and 25% close, you're at $320 per booked job before anything else. If it's a $300 service call, you worked that job for the platform.
One thing worth saying before the math continues: neither Angi nor Thumbtack publishes a flat rate card, so any specific price you see quoted online — including in this post — is either an anecdote or a labeled hypothetical. We've broken down how each platform's pricing is actually structured in Angi lead costs for electricians and Thumbtack costs in 2026. The structural takeaway from both: the platform sets the price per lead, the price moves, and the only number you can trust is the one you compute from your own invoices.
The invisible line items
- Quoting time. Each lead takes calls, maybe a site visit, and a written estimate — call it 1–2 unpaid hours. At 20 leads/month that's a 30-hour month of free work. What's your rate? That's the cost.
- No-shows and ghosts. The homeowner who books a walkthrough and evaporates burns drive time on top of quote time.
- The race tax. Shared leads reward the fastest dialer, not the best electrician — so you either interrupt paid work to chase new leads, or you lose them. Both cost money.
- Price erosion. Bidding against four competitors on every job drags your book toward the bottom of your range. A few points of margin across a year is real money.
The first line item is big enough to deserve its own post: free estimates are unpaid sales labor, and we costed a full year of them here.
A worked month
| Line | Amount |
|---|---|
| 20 leads @ $80 | −$1,600 |
| 5 booked jobs (25% close) | cost basis: $320/job |
| ~25 hrs quoting/chasing @ $100/hr opportunity cost | −$2,500 equivalent |
| True acquisition cost | ≈ $820 per booked job |
Rerun it with your real numbers — the shape survives even generous assumptions. And this is the model's working-as-intended case; the FTC documented what happens when it doesn't. (See: the HomeAdvisor order, explained.)
Track your own number: a monthly ledger you can copy
Everything above is our math. What decides whether you keep buying leads should be yours. You don't need software for this — one page, six lines, ten minutes at the end of each month. If you buy from more than one platform, run a column per platform; the side-by-side is usually the most interesting thing on the page.
| Ledger line | Where it comes from |
|---|---|
| 1. Lead spend | The platform invoice — all of it. Per-lead charges plus any subscription, ads, or "pro" fees. |
| 2. Leads received | Count every one, including the duds you disputed. Disputes you win come off line 1, not line 2. |
| 3. Jobs quoted | Every estimate you wrote, called in, or drove to from those leads. |
| 4. Jobs booked | Signed and scheduled. Not "sounded interested," not "said he'd call back." |
| 5. Revenue from those jobs | Invoiced totals once the work is done. |
| 6. Unpaid hours | Calls, drive time, walkthroughs, write-ups, follow-ups, no-shows. If you won't log daily, track one honest week and multiply by four. |
Then three formulas, in pencil at the bottom:
- Cost per booked job = line 1 ÷ line 4. This is the number the platform never shows you.
- True cost per booked job = (line 1 + line 6 × your billable rate) ÷ line 4. This is the number that decides whether the channel is worth it.
- Acquisition share = true cost ÷ (line 5 ÷ line 4). The slice of the average job that went to getting the job instead of doing it.
Once the basic version is habit, add one refinement: split the ledger by job type. Service calls, panel work, EV chargers, and remodel wiring close at different rates and gross wildly different profits, so a channel that's poison for $300 troubleshooting calls can be tolerable for $12,000 heavy-ups. Most electricians who "make leads work" are unknowingly cross-subsidizing: the big-ticket wins are quietly paying for a pile of small-ticket losses. The split ledger shows you which leads to keep buying and which task types to turn off in your platform settings this afternoon — often the single highest-return fifteen minutes in this whole post.
Run it for three consecutive months before you judge anything — one month proves nothing in either direction. A single big panel job can flatter a terrible month; one ghost-heavy stretch can slander a workable channel. Three months of the ledger gives you a trend, and the trend is the verdict. It also gives you something most contractors never have: leverage. "Your leads cost me $412 per booked job last quarter" is a very different phone call than "your leads feel expensive."
Price erosion: the cost that compounds
Every line so far is a cost you can point at. This one never appears on any invoice, which is why it does the most damage: winning a shared lead usually means winning a visible bidding contest, and the reliable way to win those is to shade your price.
Put a hypothetical number on it — clearly labeled, rerun it with yours. Say competing against three other bids shades your average winning ticket by just 5%. On $4,000 of lead-sourced revenue a month, that's $200/month — $2,400 a year — that you left on the table to beat people you never met. Shade 10% on a bigger book, say $15,000/month of lead work, and you're giving up $18,000 a year. That's a helper's wages, forfeited invisibly, in $50 and $100 slices nobody ever writes down.
And it compounds, because prices anchor. The customer who hired you at the shaded price refers you at the shaded price — "he did ours for $1,100" travels faster than any ad. Next year's book gets built on this year's discounts. Raise rates later and you're not adjusting a number, you're breaking an expectation with the exact customers most likely to refer. Lead platforms don't set your prices — but the auction they put you in does, one shaved quote at a time, and the platform's invoice will never mention it.
The defense is boring and effective: a written rate card of your own, and the discipline to lose bids below it. If a channel only produces work under your floor, that isn't a pricing problem — it's the ledger telling you what the channel is really worth. Erosion is a choice the auction pressures you toward; it stops being a cost the day you stop consenting to it, even if that means buying fewer leads.
The strongest counter-argument, taken seriously
"But leads are instant volume." It's the best argument for the model, so let's not strawman it: it's true. Nothing else fills a dead calendar by Thursday. A new shop with no referral base, an established one entering a new county, a brutal slow February — in all three cases an idle truck costs more than expensive leads, and buying work at a bad margin genuinely beats no work at a perfect one. Anyone who tells you paid leads never make sense is selling something too.
Here's where the argument breaks: it's a case for leads as a bridge, and the model quietly converts the bridge into a foundation. The spend never sunsets on its own — every month starts at zero, because the platform owns the customer, the review, and the relationship, and rents you the introduction. Ten years of buying leads builds you nothing you can stand on in year eleven. Compare that with the channels that compound — repeat customers, referrals, a network relationship — where this year's work makes next year's work cheaper. Lead spend is rent. Rent is fine short-term; nobody gets ahead paying it forever.
So the honest version of the counter-argument earns three conditions: cap it (a fixed monthly budget, not "whatever the leads cost"), track it (the ledger above, monthly, no skipping the unpaid hours), and date-stamp it (a bridge has a far bank — know what channel is supposed to replace this spend, and check quarterly whether it's growing). If you're buying leads under those three rules, we have no quarrel with you. If the spend has no cap, no ledger, and no exit, it isn't a strategy; it's a habit with an invoice. The full worked P&L of a leads month next to a restructured one is here: lead fees vs. real margin.
Why the invoices feel smaller than the math
If the true number really lands near $800 per booked job, why do so many capable electricians keep paying it for years? Because the billing is engineered — intentionally or not — to feel small. You never see an $800 invoice. You see a $75 charge on Tuesday, a $90 charge on Thursday, each one small enough to shrug at, none of them attached to the outcome it did or didn't produce. The unpaid hours never generate paper at all. And by the time doubt creeps in, you've built a profile, banked reviews, and learned the platform's quirks — sunk costs that feel like an asset and argue for one more month. That's why the ledger matters more than any argument on this page: it's the only place all the drips land in one line, next to the only denominator that counts. Feelings negotiate with a $75 charge; nobody negotiates with their own quarterly average.
What "no acquisition cost" looks like
Now price the alternative structure: a job network hands you a sold, scoped, scheduled job with materials on site and pays you a fixed number you approved in advance. Your acquisition cost is zero. Your quoting time is zero. Your collections time is zero. The network keeps a spread on the customer side — and your comparison isn't that spread, it's the $820 above.
That's the whole argument. It's also how Loadside works — and because "we pay you" earns skepticism, we wrote down exactly how we make money, spread and all. Run your own math; ours only has to beat the true number, not the sticker one.
The decision framework
Pull the last three months of ledgers and walk it in order:
- Compute true cost per booked job — lead spend plus unpaid hours at your rate, divided by jobs booked. Three-month average, not your best month.
- Compute gross profit on your average lead-sourced job — invoiced revenue minus materials and any subs, before paying yourself.
- If true cost exceeds gross profit: you are paying to work. Not "thin margin" — negative. Stop buying this channel now, before the next invoice, and restructure around channels with zero acquisition cost. There is no volume level at which losing money per job turns a profit.
- If true cost eats most of the gross profit (you're working jobs mostly for the platform's benefit): leads are a calendar bridge only. Cap the monthly spend, keep the ledger running, and put a date on the exit.
- If true cost sits comfortably below gross profit: the channel works for you today — genuinely, and some shops in some markets get here. Two checks before you relax: price erosion (is the auction quietly setting your book?), and trajectory (is the cost per booked job trending up as more contractors pile into your zip codes?). A channel can be profitable and still be a melting asset.
Wherever you land, run the same framework on the alternatives — including us. A job network's pitch isn't "free money," it's a structural trade: zero acquisition cost and zero quoting hours in exchange for the customer relationship on network jobs. If your ledger says your leads cost $800 a booked job, that trade is obvious. If it says $150, maybe it isn't. The point of this whole pillar is that you should know which one you are. If the number comes back ugly, the application takes five minutes.
And if you do nothing else from this post: start the ledger this month. Not after you quit the platform, not once things slow down — now, while the invoices are live and the hours are fresh. Every decision above gets easy once the number exists. Until it exists, the platform is the only party in the relationship doing arithmetic, and that asymmetry is the most expensive thing on this page.