What I learned building a travel agency from nothing
Not advised on. Built — the site, the booking flow, the automations, the servers underneath. Including the part that has not worked yet, because leaving that out would make the rest of it worthless.
Most people who sell marketing services have never carried the whole thing. They run the campaigns and hand the outcome to somebody else's operation. When it fails outside their slice, they are genuinely not at fault, and the problem is genuinely not fixed.
I wanted the version where no part of it is somebody else's problem. So I built a travel agency.
The question that started it
I spent four years inside Starwood Hotels. What I took from it was not industry knowledge — it was a standard: the moment a guest asks for something is explicitly somebody's responsibility, measured, and never left to whoever is nearest.
Years later, looking at how independent travel businesses handle an enquiry, the contrast was the whole opportunity. Not better photography or cheaper packages. Just treating the first sixty seconds after someone asks as though it mattered.
Infrastructure before product
I built the servers, the domains, the mail and the deployment path before there was anything to sell. That order looks backwards and I would do it again.
Not because infrastructure is more important than product — it is not — but because everything you build afterwards inherits its constraints. A business that starts on somebody else's platform makes every later decision inside a box drawn by that platform: what you can measure, what you can automate, what happens to your data if the price changes.
Owning that layer is the reason I can answer questions about my own operation that most owners cannot answer about theirs.
The site as the machine, not the brochure
The site is not there to describe the agency. It is there to catch an enquiry and route it to a human within seconds, capture what was asked, and record what happened afterwards.
Everything visible — the pages, the destinations, the writing — exists to get somebody to the point of asking. Everything invisible exists to make sure that ask does not evaporate. The second half is where almost every business I audit is losing money, and it is the half nobody photographs.
The automations that make it viable alone
A one-person operation with the response expectations of a larger one only works if the machine does the remembering.
Enquiries reach me on the device I already have in my hand, complete enough to answer immediately. Content publishes on a schedule without me being present. Follow-ups happen at intervals I decided once, not when I remember. Reports arrive whether or not I ask.
None of this is clever. It is ordinary work, done once, so that it stops needing attention. The compounding comes from it running on the days I am not there.
The part that has not worked
Here is where this article stops resembling the ones written to impress you.
The infrastructure is finished. The automations run. The content publishes daily without intervention. And organic search sends the agency a handful of visits a month — a number so small it would be dishonest to describe as a start.
I know it precisely, because I measure my own properties the way I measure a client's. The build worked. The demand has not arrived.
That is a real and extremely common condition, and it is the one nobody writes about. It is far more comfortable to publish the version where the machine was built and the customers appeared. The honest version is that those are separate problems, the second is harder than the first, and being good at building is not the same as being found.
I am including it because the alternative is a case study that reads like everyone else's, and because a consultant who only publishes the wins is telling you something about what they will do with your numbers.
What it actually taught me
Response time beats almost everything. The standard I brought from hospitality has been worth more than any optimisation I have made to a campaign, in my business and in every client account since.
Owning your infrastructure changes what you can know. Not for ideological reasons. Because when the data is yours, the questions you can ask about your business are unlimited, and when it is not, you ask the questions the platform allows.
Building is the enjoyable part and it is not the constraint. I spent longer than I should have making the operation more elegant while the actual bottleneck — nobody arriving — sat untouched. Knowing this intellectually did not stop me doing it. That is worth admitting to anyone who is about to make the same trade.
Which is, in the end, why I would rather show you something I built and operate than a slide about somebody else's results. You can open it. You can see what works. And I will tell you what does not.
Want the same look at your operation?
The acquisition audit does to your business what I do to mine: follows the money, finds where it leaks, and tells you the uncomfortable part too. Audit · $200, credited in full if we work together.