- Case study 02
- Growth site
Role
Designer · Developer
Scope
SEO · Web design · Copy
Impact
100+ new accounts a month
The context
Sales grew. The pipeline didn’t.
The sales team doubled in a year. The pipeline behind it did not. Every lead we paid for was worked within days, and the next month started empty again.
Ads were the only source, so growth meant spend. We wanted a channel we owned: long-tail SEORanking on the many low-volume, high-intent queries rather than fighting for the few competitive ones. to earn the queries, SEASearch engine advertising: buying placement on those same queries, which prices the intent before a page is built for it. to find out which ones were worth earning.
The goal
Build a pipeline we don’t rent.
The strategy
Win the questions nobody bids on.
One page for every question a farmer types. Low volume, high intent, and nobody bidding against us.
The hook
The entry point is a search box.
appli traçabilité agricole
Traçabilité automatique - Agronomie intelligente
Réduisez votre IFT de 0,25 pt/an. Simplifiez la HVE. Traçabilité automatique incluse. HYGO analyse la météo à la parcelle et recommande le bon créneau ...
One intent, one page.
The front door
Every generic query lands here. One page states the whole job, then hands the visitor onward.
651 visitors per month
38% to the app store
Own the subject
The reference on spraying: one page per registered product, built from the regulatory database.
Iterations on the catalogue
One page took 85% of the traffic. It converted worst.
The right-hand column
The sticky widget asking for the download went unanswered.
Before
- A headline with no call to action
- A button with no affordance
- The app nowhere in sight
2%
Clicks on the panel
After
- An action verb and what it delivers
- A clear, store-badged download button
- The app, front and centre
6%
Clicks on the panel
The body
The page asked once, at the bottom, where three visitors in four never go.
Both versions at full length, against how far the page is actually read. The marks are placed by the section they landed on rather than by a depth: these sheets are generated per product, and no two are the same length.
Before
- One CTA, at the foot of the page
- An unbroken run of reference text
- Nothing above the fold
Under 5%
Store clicks
After
- Five CTAs through the body
- Reference text folded away
- Two of them above the fold
13%
Store clicks
Before
After
One CTA, and it sits below every mark.
Five CTAs. Two are visible before a scroll.
One curve grades both versions. It was measured before the rebuild, which is the profile the new page had to work against. Four of the five new CTAs sit above the one the old page had.
Under 5%
Store clicks, before
13%
Store clicks, after
Visitors who left the whole sheet for the app store, by any of its routes — the panel above is one, the three banners and the foot of the page are the others. The earlier figure is a ceiling and not a measurement. The one after it is counted, and it is still the lowest of the eight pages, on more traffic than the other seven put together, against a site average of 16%.
The outcome
The next month no longer starts empty.
Every page in the row earns its own traffic, and the accounts it produces are the leads the team used to buy. Nobody bids on the questions it answers, so nothing has to be bought back the following month.
Ads never stopped. This is what was put beside them, and it is the part that keeps running when the spend does not.
8,100
Visits a month
Across the eight pages, over three months of analytics.
16%
Leave for the app store
The site's own average. The catalogue alone is 13%, on more traffic than the other seven put together.
100+
New accounts a month
A floor, not a count: consent and attribution lose about a fifth of it.
Other cases
