Lodgify case study · Head of Performance Marketing
Robert
Costart
Three problems from the brief. Three plays to answer them. All built for real, in Lodgify's own design system, using claims from their live pages.
Everyone sits in a channel silo. Budget gets tied to roles, and the weeks fill up with justifying and explaining. The team stops focusing on performance and starts focusing on answering questions. Small optimisations get defended because they're easy to explain, and the big stuff never gets tried.
Break it into a product setup, and with AI that has taken on a life of its own. We work in sprints. People bring ideas and push completely new things instead of over-optimising small details. The question becomes "what big play are we putting on the field this sprint".
My reporting job changes with it: I'm telling the business what this sprint will land and the impact it should have, not explaining weekly trends.
Specialists who ship: the whole team pushes to GitHub, AI reviews every change, and it's live on Cloudflare the same day. No channel silos, no dead budget, and one sprint question for everyone: what are you bringing that's new? This site was built exactly that way, and the three plays below all run on this engine.
I like to hit the ground running, but not by interfering. The first few weeks are for understanding: how the system works, who the people are, and when the next planning cycle lands. That's also when the quiet fixes happen, like the broken CAC definition sitting in the metric layer.
A lot of experience of what works. I'll push the things I know work and show the team a few of them, but these are ideas, not orders. The team might have even better ones, and getting them on board is the multiplier.
Depending on when I start, the calendar decides the pace. The ideas are ready either way.
These are the ideas I'd bring to that first planning cycle. And they're not theory: every one is already built and clickable below.
Stop chasing the auction with bids. Each top term graduates into its own ad group with bespoke ads, sitelinks and callouts, and every selling point gets a stripped, message-matched sub-page. Relevance raises Quality Score, Quality Score lowers CPC, right where CPCs doubled. The sub-pages below become the sitelinks of the ad and the retargeting messages after the click.
A/B testing built into the landing system itself, no new tools needed. Every sub-page above doubles as a test cell: message variants in, winners promoted, losers quietly retired. Then the same system funds the big test. We repurpose the near-zero display spend into a steady assurance layer around new trialists, with a 50/50 holdout reading the true lift on trial-to-paid. Prove it works first and optimise the cost second. Even a null result frees €360k a year.
Landlords start their hosting journey searching like travellers, things like “Barcelona rentals” and “London flats”. Lodgify has zero city pages today but owns the data to build them: real inventory density in every market, plus thousands of customer websites it builds and hosts, each linking back. Location pages catch host journeys right at the start. It's a compounding channel that cuts the 67% dependence on paid Search.
This is the answer to the brief's geo and channel question. Stop hand-allocating budgets in meetings and let value-based bidding do the moving.
That's the brief, in Lodgify's own words. Everything on this page is live and clickable, and it was all built in days on the operating model above. That's the point.
