BUILDING A GROWTH ENGINE FROM ABSOLUTE ZERO

This is not a case study written after the fact by someone who had a plan and executed it cleanly. Decisions were made without full information. Problems were found after they'd already been running for weeks. A few things had to be rebuilt from scratch mid-way through.

The numbers are real. The problems are real. And some of them had nothing to do with advertising — they showed up in what happened after a lead came in, which turned out to matter just as much as how the lead got there.

This is the full story, from the first day.

DAY ONE. EVERYTHING IS MISSING.

The concept was sharp: connect patients in underserved markets — primarily Nigeria and wider Africa — with specialist healthcare in India. Haematology. Bone marrow transplants. Oncology. Complex surgeries where the gap between what's available locally and what's needed is significant.

The model was simple: patients seeking specialist medical care — primarily from Nigeria and across Africa — get connected to hospitals and doctors in India. Everything in between, from the initial consultation to travel and appointments, is handled end to end.

Good model. Real need. Experienced founders. And then I logged into the systems for the first time.

I want to be precise about what "no digital presence" actually means in practice, because it's easy to read that and think I'm exaggerating. There was a domain. There was a placeholder — a holding page — but nothing that could receive traffic, convert a visitor into a lead, or tell me anything about who was finding us, how they were finding us, or what they did when they got there. No pixels, no tags, no goals, no events. If someone in Lagos had typed our company name into Google that week and landed somewhere, we would have had no record of it happening.

For a company targeting patients in Nigeria and Africa who are desperately researching treatment options — often at two in the morning, on a phone, terrified about a diagnosis they received that week — that was a problem that needed to be solved before anything else.

I spent the first week doing nothing but mapping what needed to exist before I could build anything useful. Website. Analytics. Search Console. Tag Manager. Conversion tracking. Account structure. Keyword research. Landing pages. Campaigns last.

GETTING THE WEBSITE RIGHT. BEFORE THE ADS, BEFORE EVERYTHING.

The instinct in this situation is to get the ads running first and sort the website out later. I understood the pressure — the founders wanted leads, and they wanted them soon. But sending paid traffic to a placeholder page burns budget and proves nothing. I made the case for sequencing it correctly, and the first month was infrastructure.

The company had an external IT team who handled the technical build. My role was to define everything that would affect ad performance and conversion: the site architecture, the brand system, the page structure, the copy, and critically — the performance standards the site had to meet before we pointed paid traffic at it. Mobile page load speed is a direct input into Google's Quality Score via the "Landing Page Experience" component. A slow site means higher CPCs in every single auction, indefinitely. So those performance requirements weren't aesthetic preferences — they were ad account constraints I set upfront and worked with the IT team to hit.

The brand direction I defined was built around the nature of the service. Patients arriving at this site have usually already been through something frightening. Deep evergreen tones, restrained layout, clean typography — nothing that felt clinical or transactional. The content on every page was structured around the questions a patient or family member actually types at two in the morning, not what a company brochure would say about itself.

Eight core pages: home, services, about, the doctor consultation pathway, condition-specific pages for haematology and oncology, a destinations page for the India hospital network, and a contact page built around a single focused lead form. Site went live in late July — functional, reasonably fast on mobile, and with tracking in place from day one.

SETTING UP THE MEASUREMENT STACK. THE UNGLAMOROUS PART NOBODY TALKS ABOUT.

Nobody writes case studies about setting up Google Search Console. Nobody posts about linking GA4 to Tag Manager on a Tuesday afternoon. It doesn't have a dramatic number attached to it. But I want to write about it here, because if you skip this step — or rush it — you spend months making decisions based on data you can't actually trust. I've seen what that looks like. I wasn't going to let it happen here.

THE RESEARCH PHASE. UNDERSTANDING WHAT WE WERE ACTUALLY SELLING.

Campaigns went live on August 15th. But before I touched the ads interface, I spent three weeks in research that most people skip. I want to walk through what that looked like, because it shaped every structural decision I made later — including some decisions I had to revisit when the data came back and told me I'd been wrong.

THE FIRST CAMPAIGNS GO LIVE.

August 15th. Six campaigns structured around intent tiers — each with its own budget, its own tightly themed ad groups, its own corresponding landing page. Responsive Search Ads across every group, fifteen headlines and four descriptions each, written to the specific search intent of each keyword cluster rather than generic copy.

Extensions built out fully from day one: callouts for the free consultation and end-to-end coordination, sitelinks to specific condition pages rather than the homepage, structured snippets listing actual procedures. The goal was to make the ad useful before someone even clicked — so the clicks we paid for came from people who already understood what we offered.

The first few weeks were deliberately treated as a data collection phase. Before smart bidding can work efficiently, the account needs a minimum volume of conversion signal. Running manual CPC initially and letting the campaigns accumulate real search term data was the plan — not to hit targets yet, but to have something honest to audit. After about six weeks, I ran that audit. The data had found some things worth looking at.

THE SIX-WEEK AUDIT. WHAT THE DATA ACTUALLY SHOWED.

The impression share report was the first thing that flagged. We were losing a significant chunk of available impressions — and the reason wasn't budget. Google was choosing not to show our ads as often as it could because the ads-to-landing-page experience wasn't scoring well enough in its auction logic. That pulled me into a full account audit: Quality Score breakdown by keyword, Page Speed Insights on every landing page, search term analysis, campaign structure review.

A Quality Score average of 5.74 means you're losing auctions to competitors who have the same bid as you but cleaner accounts. You're paying more per click than you should. You're showing less often than you could. And it's a compounding problem — lower Quality Score means worse Ad Rank, which means lower visibility, which means fewer clicks, which means slower conversion data for the algorithm, which means less efficient optimisation. Everything gets worse together, slowly and invisibly.

Running Page Speed Insights on the landing pages confirmed the source. Mobile LCP — the time it takes for the main visible content to appear on a mobile device — was at 5.4 seconds. Google's threshold for "Good" is 2.5 seconds. More than double, on mobile, in markets where most of the audience browses on a phone on a 4G connection.

THE REBUILD. FIXING THINGS IN THE RIGHT ORDER.

When you have multiple compounding problems in an ads account, the instinct is to start with the most visible one — the bids, the budget allocation — because those feel like immediate levers. That would have been the wrong move. The problems were chained. The landing page was suppressing Quality Score. Quality Score was suppressing Ad Rank. Ad Rank was suppressing impression share. Until the landing page was fixed, nothing else would hold. You start at the root, not the branch.

WHEN THE LEADS CAME IN, A DIFFERENT SET OF PROBLEMS APPEARED.

Once the ads started working, it became clear that the ads account was only one part of the system. The lead comes in, the form gets submitted, the CPA looks clean in the dashboard — and then you find out what actually happens next.

I started sitting with the patient coordinators regularly. Not to report back to management — I wanted to understand the pipeline from their side, because I was optimising based on form submissions and had no visibility into what happened after a lead landed.

BUILDING THE TRACKING LAYER FOR THE SALES PIPELINE.

I could fix the ads. I could fix the landing pages. I could sit with the coordinators and help redesign how they engaged with leads. But none of that learning was being captured anywhere in a form that persisted. Every conversation about lead quality happened verbally and then dissolved. The following week, decisions were being made on memory and gut feel rather than patterns visible across actual data.

WHERE IT STANDS. RESULTS ACROSS ALL THREE PHASES.

Here is where everything stands.

The $29 figure is cost per form submission, not cost per patient. Total spend across the full campaign period was $16,465, generating 569 submissions — that maths holds. But a form submission is the entry point to the pipeline. The real cost per consultation, and per case forwarded to a hospital, are higher — and we can now calculate both for the first time through the CRM. That work is ongoing.

What the number doesn't capture is the full scope of what was built: from a patient finding the company through a Google search late at night, to submitting a form on a mobile phone, to a consultation with a doctor, to their travel and hospital appointment being coordinated end to end. That pipeline didn't exist when I joined. It does now.