Facebook and Meta Ads for B2B SaaS

Why Meta ads usually fail for SaaS

The SaaS version of the failure has a particular shape, and it starts with LinkedIn.

You ran LinkedIn ads first, because that’s where the job titles are. They worked, at $70 to $100 per thousand impressions, and finance asked why the CAC was what it was. Somebody said Meta is a fifth of the price. So you took the same whitepaper and the same lead form to Meta, targeted “B2B software” and “marketing directors,” and got 300 downloads for the cost of 40 on LinkedIn. Then sales called the downloads. Students, consultants, people at 8-person companies, and a few real prospects who didn’t remember downloading anything. Four demos. Meta went in the “doesn’t work for us” column.

Three things went wrong, and none of them was the channel.

The audience was a guess. Meta removed most detailed B2B targeting in 2024. “Marketing directors” on Meta is people who put that in a profile field, which is a small and strange slice of the actual marketing directors in the country. The real ones are on Meta, in the evening, on their phone, and there’s no interest that finds them.

The offer was a download. A whitepaper is the right offer on LinkedIn, where someone is in a work mindset. On Meta at 9pm it’s a thing people tap and forget. The people who tap it are not the people who’ll take a demo.

Nobody worked the leads. Three hundred form fills went to the CRM, sales called the first fifty, the ratio was bad, and they stopped. If the real prospects were fills 61, 140, and 212, nobody found them.

The only thing we do with Meta ads

We book B2B sales calls. For a SaaS company that means every audience, every ad, every form, and every dollar of spend exists to put someone who owns the problem your product solves onto your AE’s calendar, and the number at the top of your daily log is demos booked, not MQLs and not downloads.

Where Meta fits for SaaS, in order

There are three uses, and they work in this order. Most SaaS companies start with the third one, which is why it fails.

Retargeting comes first. A B2B SaaS company has something an MSP or a machine shop doesn’t: a website with real traffic, a pixel, and a list of people who signed up for something. Anyone who visited your pricing page, watched a product video, opened a sales email, or started a trial and stalled is a warm audience Meta can reach for a fraction of what LinkedIn charges. This is the cheapest demo you’ll ever book, and most SaaS companies aren’t doing it.

Signal audiences come second. A signal audience is a Meta custom audience made of named companies that just showed a public reason to buy, matched by contact and company domain, and rebuilt every day. For SaaS, the signals are a job posting for the role your product serves, a funding round or a new VP, a stack change you can see in job posts or on their site, a competitor’s bad reviews ahead of renewal, and expansion into a new segment. We watch all five across your ICP every day. Business emails match worse on Meta than personal ones, so we add company domains and page engagers, and the match rate lands in a usable place for mid-market. For enterprise accounts it’s thinner, and we’ll say so on the fit call.

We don’t run cold interest audiences at all. If your agency is targeting “SaaS” or “B2B marketing” as an interest, that’s where the students come from.

Then there’s the lookalike, built from your closed-won customers in the CRM, not from site visitors, so Meta models the companies that actually pay you. It’s the audience that finds the companies the signals missed.

What the ads say

The creative follows the outbound copy. Short, about the buyer’s situation, and almost never a product tour.

Your founder or your best AE on camera for 30 to 45 seconds, about one signal. “If you just posted a RevOps Manager role after a round, here’s what we’ve seen happen to forecasting between 30 and 60 reps.” One signal per video. The company that posted the role sees that one. The company that just adopted HubSpot sees a different one, about the integration.

A screenshot of a real reply. A VP writing back “yeah, let’s talk, when’s good” with the name blurred. On Meta this beats designed creative every time we’ve tested it for SaaS, because it’s the only thing in the feed that looks like a person.

A product screenshot with one line. We don’t run tours or feature lists, just the one screen that answers the question the signal raised, with a sentence of text. The forecasting view, for the company that posted the RevOps role.

A case study card. One number, one company, one sentence. Gainsight’s is the first one we run.

We run both a lead form and a landing page with the calendar on it, and we test which books more demos for your ICP. The lead form offer is a demo, not a download. A form fill isn’t the result. Someone on our team emails and calls every fill the same day, sorts the buyers from the students and the consultants, and books the demo on your AE’s calendar. What we report is the demo.

How we measure it

We report reach, clicks, and cost per click because they’re the only way to tell whether the audience is being served at all. They’re not the result.

The result is cost per booked demo and cost per held demo, measured off your AEs’ calendar and not off the pixel. Every day you get spend, form fills, demos booked from those fills by our team, demos booked from the landing page, and, if you run our email too, demos booked by email from companies that also saw the ads. Your AE can reject any demo and it comes off the count, same as on the email programs.

A SaaS test needs six to eight weeks and somewhere around $3,000 to $8,000 a month in spend to produce a number worth judging. Retargeting is cheap and fast and usually books demos in the first two weeks. Signal audiences take longer to fill and cost more per reach, and they’re where the net-new pipeline comes from.

How it runs

Week one we set up or audit your ad account and pixel, agree on the ICP and what counts as a demo, build the retargeting audience from your pixel and your email list, build the signal audience from our monitoring, build the lookalike from closed-won, and script the first video.

Week two we shoot or collect the creative. One video per signal you want to run, two screenshot ads, one product screenshot, and the Gainsight card. Ads are live by the end of the week, retargeting first.

Weeks three to eight the audiences and creatives rotate against each other. Every form fill gets an email and a call from our team the same day. Every morning we look at yesterday’s booked demos against the target. If we hit it, spend goes up. If we missed it, it’s one of four things: the audience, the spend, the creative and offer, or the follow-up. We find which, fix it, and that’s what runs today. You can follow it in the log.

The ad account stays in your name the whole time.