How to scale B2B SaaS paid ads profitably: an interview with performance marketing leader Kenan Choufi

The biggest shift is buyers doing their research in ChatGPT and Perplexity before they ever click a paid ad, so by the time they hit your landing page they're already comparing you to competitors. Paid still works, but it's now competing for attention earlier in the funnel, not just at the search box.

Most of the waste now comes from platforms pushing more control away from advertisers — like broad match by default, AI Max on Google Ads, or Advantage+ on Meta. Under poor management, irrelevant queries or audiences eat up the budget.

PMax helps once you feed it offline conversion imports tied to actual paid revenue instead of just a sign up or form fill. Without that signal, it optimizes toward whoever converts easiest, which for SaaS usually means free trial signups that never become paying customers. Brand exclusions are a must too, otherwise PMax will just capture your own brand search and take credit for it. I have yet to see Advantage+ do anything positive on Meta.

LinkedIn earns its spot for high-ticket ABM where the audience targeting justifies the cost, and Google search still wins for capturing people who already know what they want. Meta works more as a lower-cost reach play — less precise on hitting actual decision makers, but a cheap way to get in front of the right person at scale.

The bigger shift is that buyer research increasingly starts in AI tools, so getting cited or recommended there is becoming its own channel to manage. More businesses run daily tasks through Claude or ChatGPT now, so the instinct to open a search engine or scroll social for a solution is fading fast.

Companies that scale profitably don't hand full control to automated strategies in the ad platforms until they have enough conversion volume to actually trust it — they earn their way into it with real data first. They also pull in offline conversion tracking and other data sources before making budget decisions, and they check LTV:CAC by channel before scaling spend further. Ad platforms will always attribute more credit to themselves to justify more spend, so that self-reported data can't be the only input.

The companies that end up spending into a rising CAC are the ones who lean on automation too early and keep increasing the budget without checking those signals, mistaking more spend for more pipeline until the math stops working.

Airtop's Mark is a product I'm close to, having advised on the development, and it's genuinely useful for paid teams building campaigns from scratch — think of it as an assistant that handles the tedious build work that goes into Google Ads so you can focus more on strategy and optimization. Claude is my go-to for creative briefs and refining strategy. The bigger point is paid teams need to invest more time into using AI to automate the busywork, so performance marketers can focus on the strategy and the levers that actually move ROI.

The number that actually matters is revenue-per-dollar-spent by campaign and keyword, tracked through your CRM all the way to closed-won — not lead volume or cost-per-lead. I've had campaigns with a worse cost-per-lead outperform a cheaper one because those leads booked calls at a higher rate and closed bigger deals. Of course the simple lever-pulling metrics like CTR and CVR are still useful, but only as diagnostics for where the funnel breaks, not as the headline number you report up.

Over the next two to three years, AI-driven discovery becomes a real acquisition channel alongside search and social, and teams that aren't tracking referral traffic from AI tools will be flying blind. The prep work now is building attribution and making sure your brand shows up well when these tools get asked about your category. I already see this in GA4 for larger, credible companies — tools like ChatGPT send thousands of visitors to their sites every month with zero ad spend behind it.