A healthy-looking cost-per-lead can be a symptom of an account getting worse, not better. In B2B SaaS, the cheap leads are usually the expensive ones — and cost-per-lead is the wrong scoreboard to be cheering. The number that actually matters is cost-per-SQL (or cost-per-opportunity), not cost-per-lead. So if your CPL looks great but pipeline is flat, the problem usually isn’t your budget or your bids. It’s that you’re optimizing toward the cheapest possible lead, and the cheapest lead is almost never your buyer. Here’s how to tell whether your “good” CPL is real, and what to watch instead.
The hard part isn’t seeing the problem. It’s accepting a more expensive CPL.
By the time someone calls us, they usually already suspect something’s off. We don’t get many audits from people who think their account is great. One marketing lead put it bluntly: they’d fired their last performance marketer “because we felt like he was wasting money” — and they’d stopped trusting paid marketing in general, mostly because nobody could ever show them the results in a way that tied to revenue.
So the data isn’t the hard part. The hard part is the mind-flip: getting comfortable with a higher cost-per-lead because it’s buying you better pipeline. That feels wrong. Cheaper looks like winning. But as we cover in our take on what actually drives B2B SaaS pipeline, the easiest thing for any ad platform to get you is a cheap, unqualified lead. A falling CPL can just mean the algorithm got better at finding people who’ll never buy.
Why a cheap CPL is usually a symptom, not a win
When a non-brand campaign is posting suspiciously cheap leads, it’s almost always one of two things.
1. Brand terms are hiding inside it. Your own brand name converts at roughly a fifth of the cost of everything else, so when brand searches sneak into a non-brand campaign, they drag the whole CPL down and make it look like a star. The problem: you learn nothing. A bottom-funnel campaign that’s quietly converting on your brand name isn’t telling you whether your actual targeting works. The fix is to negate those terms and pull them into a dedicated brand campaign, where you control them and can finally see how the non-brand campaign performs on its own.
2. The search terms under the hood are junk. This is the one that surprises people. You can target a keyword that looks hyper-specific — “construction management software for electricians” — click into the search terms report, and find you’re actually showing for “electricians,” “house software,” “renovation tracking.” A practice-management product for accountants bids on “tax software” and gets a flood of consumers trying to file their own taxes during tax season. The keyword looks targeted. The traffic is B2C.
There’s a fast confirmation: go look at the conversions in the CRM. If the cheap leads are mostly Gmail and Yahoo addresses, you’ve got your answer. Businesses buying real software tend not to sign up with their personal email.
How to tell if your “good” CPL is actually real
You don’t need full CRM access to get suspicious. Here’s the order I actually check things.
Start with the search terms. For B2B software that costs real money — the kind a business has to approve — one-word search terms are a red flag on their face. I find it genuinely hard to believe that “tax software” produces good leads for a practice-management platform; that’s a consumer doing their taxes 99 times out of 100. Two minutes in the search terms report tells you more than the CPL ever will.
Be fair before you condemn it. One caveat: check the attribution lens before you write the channel off. The same spend can look dead or doubling depending on the model. One account’s search ads showed almost no revenue on first-touch attribution — and roughly double the spend once you looked at it through a multi-touch model. So confirm it’s really the leads that are bad, not just a reporting view making a working channel look worse than it is. (And don’t let it cut the other way either — a flattering multi-touch number doesn’t excuse a pile of Gmail signups.)
Then rough-math the funnel. Ask how many opportunities they generate a month, or customers a year, and hold it against the spend. You’ll usually spot the drop-off immediately — leads pouring in, opportunities barely moving.
Then, if it exists, look at the real number: cost-per-opportunity or cost-per-SQL. Honestly, only about one in five accounts I audit even has this set up. If yours doesn’t, that’s fine — the search-terms read and the rough funnel math are your proxy, and step one is getting the measurement live. (That’s exactly what offline conversion tracking is for — sending your real down-funnel events back so you can finally see cost-per-SQL by campaign and keyword.)
Stop optimizing to CPL. Optimize to cost-per-SQL.
Here’s the reframe that fixes the whole thing. PPC’s job isn’t to produce leads — it’s to produce qualified, ICP-fit leads and hand them to sales at the demo (or, for a PLG motion, at the trial). The close comes from the right person showing up under the right premise, a strong offer, and the sales process. So the way to judge an account, or a single keyword, is its cost-per-SQL, valued against what an SQL is worth to you — which you get from your SQL-to-close rate on your other channels.
One thing that trips people up: PPC tends to bring in bigger deals. So it can run a slightly lower close rate than your other channels and still win, because the deal size makes up for it. If you judge it on raw lead count or even close rate alone, you’ll underrate it.
The cleanest example I’ve got: a construction-safety SaaS spending around $50k a month. In the year before our audit, that produced about 225 leads — and roughly 10 opportunities. A 5% lead-to-opportunity rate. The leads looked cheap and plentiful; the demos were coming in at $100–300 each. But their ACV was $40,000–70,000. A demo that cheap against a deal that big isn’t a bargain — it’s an alarm. The account was optimizing beautifully for the wrong thing.
The bottom line
Cost-per-lead is a diagnostic, not a goal. A low one tells you almost nothing on its own, and the accounts that quietly waste the most money are often the ones with the best-looking CPL — because cheap junk hides in plain sight. Check the search terms, confirm what’s actually converting, and move your scoreboard to cost-per-SQL. Tighten the targeting, feed the platforms real pipeline instead of form fills, and let the CPL land wherever it lands.
Want to know whether your account is optimizing for leads or pipeline? That’s the first thing our free PPC audit digs into — the search-term bleed, the brand-term mirage, and what your “good” CPL is actually buying.