-
Looking at your numbers, the surface metrics don’t actually look bad. A CTR of almost 4% is respectable, and a $1.53 CPC for web development terms is on the cheaper side. Even the reported conversion rate of 12% is unusually high compared to what most service industries get. The real issue shows up when you compare those conversions to revenue. Bringing in $7-8K from $15.9K spend means each conversion was only worth about $6, which is nowhere near sustainable for web development projects. That suggests your “conversions” weren’t tied to real revenue events but likely to soft actions like form fills, call clicks, or site events that don’t actually mean a client was closed.
The second issue is targeting. Running generic campaigns like “web development” or handing everything to Performance Max is a fast way to get low-quality leads. People searching broad terms are often price shopping, just curious, or running projects too small to be worth your time. Google’s algorithms will happily optimize for cheap, easy conversions if you tell it to, which means you end up with a big pile of junk leads instead of qualified prospects.
On top of that, even if some good leads slipped through, the lack of a tight sales process probably hurt ROI further. With a spend that large, you need instant follow-up, filtering, and scheduling so you don’t lose the handful of prospects who could actually pay. Otherwise, most of the value vanishes after the click.
If I were reworking this, I’d start by redefining what counts as a conversion. A lead should only register if it’s a booked consultation or a qualified inquiry with budget and timeline, not just anyone filling a form. From there, focus on niche search terms instead of broad “web dev” ones such as “Shopify to WooCommerce migration” or “custom SaaS MVP agency.” Those bring in people with a clear intent to spend. Finally, make sure your funnel and follow-up process are bulletproof before scaling spend. Otherwise, you’ll just keep paying for activity instead of paying for actual revenue.
