What 443 leads at $297 CPL taught me about industrial B2B

$297 a lead.

That number gets reactions. Agency folks see it in a deck and roll their eyes. SaaS marketers tell me I'm doing it wrong. A Google rep, on an actual call, once asked if I'd considered "expanding match types to bring that down."

No.

Here's the thing about $297. The product is a MECCO laser marker. We're not selling a $99 SaaS seat. We're selling a multi-thousand-dollar piece of industrial equipment that bolts into a manufacturing line and runs around the clock for the next decade. The buyer is a manufacturing engineer or a procurement lead at a company that puts permanent identification onto metal parts for a living. There might be 3,000 companies in North America that even need this thing. Sales cycle runs 90 days on the short end, 9 months on the long end. And when one of those leads closes, the deal makes $297 look like a rounding error.

Over 16 months I generated 443 of those leads. Same brand, same vertical. Stable CPL. Stable volume. Stable close rate.

Here's exactly how.

The category is brutal in ways most marketers never see

The first thing you learn running paid media for industrial marking is that the category is contested by people who have nothing to do with your buyer.

"Laser marker" looks like a clean keyword. It isn't. Search it. You get DIY hobbyists, jewelry engravers, people looking to mark wedding rings, Etsy sellers, college students writing papers, kids on YouTube who want to engrave their PlayStation. Not one of them is buying a $40K industrial fiber laser.

That contamination is what kills B2B accounts. CPCs get bid up by hobbyists with cheap intent. Conversion rates collapse because half your traffic is showing up for the wrong reason. And Smart Bidding, which is constantly hunting for "easy wins," will absolutely route your money toward the cheap stuff if you let it. So it does. So I stopped letting it.

What we actually did

Three big moves. None of them are clever. All of them are things a Google account rep will tell you not to do.

1. Killed broad match. Completely.

Not "reduced." Not "tested down." Killed it. Every campaign in the account got rebuilt to use exact match, and in select expansion campaigns, phrase match. Broad was off across the board.

I know what the rep playbook says here. They'll tell you broad + Smart Bidding is the new normal, that the algorithm has gotten smart enough to find conversions you'd never think of. Maybe that's true if you're selling subscription software at scale. In industrial B2B, with a few thousand buyers in the country and 9-month sales cycles, the algorithm doesn't have enough signal to do anything but hallucinate. So it does. It maps your $40K laser system to "engrave my dog's collar" and bills you $11 a click for the privilege.

2. Built out exact match like it was a religion.

We mined search term reports weekly for two purposes. New exact match keywords (every legitimate variation of intent we hadn't already captured) and new negatives. The keyword list got huge. Hundreds of exact match terms broken out by application: dot peen, fiber laser, UV laser, CO2 laser. By industry: aerospace, automotive, medical device, firearms. By part type and metal type.

This isn't elegant. It's not what gets demoed at Google Marketing Live. It looks like a brick wall of low-volume keywords, most of which trigger maybe three impressions a month. That's the point. Three impressions from a procurement engineer at a tier-1 auto supplier are worth more than 30,000 impressions from people googling "laser cutter for sale."

3. Optimized to Salesforce stages, not web conversions.

This is the part most accounts get wrong, and it's the part doing the heavy lifting. We were not optimizing toward form fills. Form fills are noise. We were optimizing toward Salesforce opportunity stages, pushed back into Google Ads through a custom pipeline.

Practically that meant SQL became part of my job. We mapped click IDs to Salesforce opportunities, watched stage progression, pushed delayed conversion signals back into Ads weeks (sometimes months) after the original click. The bidding model wasn't optimizing for the cheap conversion anymore. It was optimizing for the conversion that became real pipeline.

Once you do that, the whole system reweights. Junk traffic stops looking attractive to the algorithm. The "low search volume" terms that actually close deals start getting fed budget. CPL goes up on paper. Pipeline gets healthier in reality. That trade is the entire game.

"443 leads in 16 months is too low"

I get this one a lot.

If you're benchmarking against B2C, sure. If you're benchmarking against ungated SaaS lead magnets, sure. If you're benchmarking against industrial equipment that closes at six figures with a 9-month cycle, 443 SQL-grade leads is a serious year and a half.

Volume in B2B is usually a tell. If a $40K laser system account is generating 5,000 "leads" a year at a $40 CPL, somebody is either counting whitepaper downloads or buying garbage traffic. Probably both. The sales team knows. They've just stopped saying anything about it because nobody wants to hear it.

I'd rather defend 443 closed-loop leads in a board meeting than 5,000 MQLs that converted to nothing. One of those conversations ends with a budget increase. The other ends with a layoff.

The blueprint, if you want to copy it

If your account looks like ours (low absolute volume, high AOV, long sales cycle, heavy consumer overlap on category terms), this is the order I'd run it in.

1. Turn off broad match for one full week. Don't reduce. Turn it off. See what happens to volume vs. close rate. In every B2B account I've done this in, total spend drops, lead count drops, and pipeline either stays flat or goes up.

2. Build exact match by application, not by category. "Laser marker" is a category. "Fiber laser marker for aerospace serialization" is an application. The application is the keyword. Build out every application you can think of, even the ones with one search a month. Especially those.

3. Build the negative list like it's the product. Mine search terms weekly. Categorize negatives by reason: consumer, hobbyist, academic, employment, used/cheap, wrong industry. Apply them at the account level so they protect everything you launch in the future too.

4. Get offline conversions wired up before you do anything else. If you're optimizing to form fills, you're optimizing to noise. Pull Salesforce stage data into Google Ads via the API. If your team can't, that's the project for this quarter. Nothing else matters as much.

5. Stop reporting CPL as a primary metric. Report CPSQL, pipeline, and closed-won. CPL is the metric agencies use to look good. CPSQL is the metric that doesn't lie.

That's the whole thing. 443 leads, $297 a piece, 16 months. No magic. No clever tricks. Just refusing to do the things every Google rep and most agencies will tell you to do.

If you're running a niche industrial account and your CPL looks suspiciously low right now, ask your sales team how many of those leads they actually wanted. The answer tells you everything.