Enterprise AI platform · B2B
Case study · Google Ads + conversion tracking · B2B
Client confidential · Full case study on request
Search couldn't scale. So I built the demand, and the tracking to count it.
Google Ads Search, GA4 and GTM conversion tracking, and organic LinkedIn for an enterprise AI platform sold through ERP partners.
42.1% CTR
on the OCR vs Document AI post
9 tagsin the GA4 and GTM build that separates idle clicks from real enquiries
From 0LinkedIn impressions, after a silent year
20.5%average click-through on 5 organic posts
Nonepaid promotion behind those results
Get ERP partners to notice a practical AI platform.
The platform (document AI, agentic AI, knowledge AI and computer vision) is sold through ERP channel partners. I ran every digital touchpoint:
Google Ads SearchWebsiteGA4 + GTMLinkedIn company pageFounder profile
02 · Search
I ran Google Ads Search to catch partners already looking.
Structure by intentCampaigns and ad groups mapped to what partners search for.
Keywords and match typesPhrase and exact for intent, negatives to cut waste.
Responsive search adsPartner-led messaging, tested headline by headline.
Search terms reviewSo spend followed real queries, not guesses.
The insight
Search can only catch demand that already exists. Almost nobody searches for "Document AI for ERP partners" yet.
So the job became two: build the tracking to count every enquiry, then create the demand where partners already spend time.
What a Search enquiry costs
Business services: benchmark cost per clickthe price of one click
× 6.64% of clicks becoming enquiriesabout 15 clicks per enquiry
10 enquiries a month needsabout 150 matching searches
Price was workable. Search volume was the limit.
At benchmark prices, an enquiry is cheap next to an enterprise partner deal. But 10 enquiries need about 150 matching searches a month, and for a brand-new category those searches barely exist yet. More budget can't buy searches nobody makes. So I moved the effort from paid search to building demand on LinkedIn.
03 · The tracking
Before buying more clicks, I built the tracking to count them.
Tracking
A GA4 property and GTM container, built from a spec
9tags
9triggers
10variables
Separates idle clicks from confirmed enquiries, qualified partner conversations and first sales.
Mobile
Found a home page 664px wide on a 390px phone
Plus a 21,861px scroll, and 0 of 16 headings naming partners. 14 prioritised fixes.
SEO and messaging
All 11 pages audited, rewritten for partners
New titles and descriptions, benefit-led copy, a partner journey map and topic clusters for content.
11 pagesJourney mapTopic clusters
The LinkedIn page had been silent for a whole year.
0posts
0impressions
05 · The fix
A 12-post education campaign, built around what partners worry about.
Brand and trustAI adoptionProduct capabilitiesPartnershipsEducation
Backed by a founder amplification plan and a reply playbook for target partners.
06 · The turning point
Every post woke the page up.
3 days before3 days after
07 · The insight
Teaching beat telling. The explainer earned twice the clicks.
45.9%engagement rate on the explainer carousel
So the next plan leads with educationThree cadence options for the client: 8 posts, 6 posts plus 2 carousels, or 12 posts a month.
What this says about how I work
I build tracking before trafficso every click can be counted.
I match the channel to the demandcapture it on search, create it on LinkedIn.
I let the best post write the plandata picks the next content, not taste.