Server-side tracking
Collection on the customer's own first-party domain, so conversions survive declined cookies, ad blockers and iOS privacy restrictions.
[ CASE STUDY ]
LEADJOURNEY.IO
Marketing attribution for businesses that sell to people rather than to carts — and the engineering problem underneath it.
- ROLE
- Head of Development & Product Owner
- TIMEFRAME
- Oct 2025 — Present
- PRODUCT
- Marketing attribution SaaS
- MARKET
- Agencies, B2B and lead generation
WHAT IT IS
LeadJourney is an attribution platform for lead generation — not e-commerce. The difference matters more than it sounds. In e-commerce the purchase happens in the browser, seconds after the click, where the pixel can still see it. In lead generation the purchase is a phone call three weeks later, closed by a salesperson, recorded in a CRM, and completely invisible to the ad platform that paid for it.
That gap is the product. It pulls ad platforms, organic and AI search, social, offline touchpoints and the CRM into one place, then attributes revenue back to the campaign that actually caused it — every touchpoint along the way, not just the last click.
The platform's own published figures: 95%+ tracking accuracy, more than fifty integrations, over a hundred marketing teams, and upwards of a million in tracked ad spend.
95%+
TRACKING ACCURACY
50+
INTEGRATIONS
100+
MARKETING TEAMS
21 MIN
AVERAGE SETUP
WHAT THE PLATFORM DOES
The surface a customer sees, and the reason the problem underneath it is worth solving properly.
Multi-touch attribution
Credit for every touchpoint that influenced a deal, rather than whichever one happened to come last before the form was filled.
Offline conversions
Calls, meetings and deals closed by a salesperson, matched back to the campaign that started them weeks earlier.
AI search attribution
Leads arriving from ChatGPT, Perplexity, Claude, Gemini and Copilot, tracked as a channel of their own. New enough that most of the market has no answer for it yet.
Atlas, the AI analyst
Questions asked in plain language against campaign, site and CRM data, answered with a recommendation rather than another chart to interpret.
Custom report builder
Reports assembled from raw event data, so an agency is not limited to the dashboards somebody else imagined for them.
THE ENGINEERING PROBLEM
Attribution looks like a reporting feature from outside. From inside it is an ingestion and identity problem with a dashboard on the end.
Tracking that survives the browser
Cookies get declined, ad blockers strip scripts, and iOS drops what is left. Browser-based tracking runs at sixty to seventy per cent and quietly loses the difference. Collecting server-side on the customer's own first-party domain is the only way to close that gap — and it has to be GDPR-safe by construction rather than by policy, because the customers are European and the regulator is not theoretical.
Fifty integrations, each broken differently
Google, Meta, LinkedIn, TikTok and Bing on one side. HubSpot, Salesforce, Pipedrive, Close and GoHighLevel on the other. Every one has its own auth, its own rate limits, its own definition of a conversion, and its own outages. Writing the client is the easy half. The hard half is behaving correctly when one of the fifty is wrong, slow, or lying.
Attribution across weeks and devices
A lead clicks on a phone, comes back on a laptop, books a call, and signs a month later. Joining that into a single journey is identity resolution, and every touchpoint in between has to be stored and credited rather than discarded in favour of whatever came last. Multi-touch attribution is a data problem long before it is a reporting one.
WHAT THE ROLE INVOLVES
Product ownership
Roadmap and prioritisation for a platform where most feature requests are a real customer's blocked integration. Half the job is deciding what is worth building and what is a support answer.
Architecture
Technical direction end to end, from event ingestion through to reporting — and specifically the decisions that are expensive to reverse once customers depend on them.
Delivery
Leading internal and external teams, shipping to customers who are running live ad budgets against the numbers on the screen.
Reliability
An attribution platform that drops events is worse than no attribution platform, because the customer trusts the wrong number instead of knowing they lack one. Reliability is not a quality bar here. It is the product.
WHAT THE JOB IS
Two halves, roughly equal in weight.
The first is deciding what gets built. On a platform with fifty integrations and customers running live ad budgets against the output, nearly every request is legitimate and most of them cannot happen this quarter. Prioritisation stops being a framework exercise and becomes a plain question: which customer stays blocked, and for how long.
The second is making sure what ships keeps working. Owning architecture end to end means owning the decisions that are cheap today and expensive in eighteen months — how events are stored, where the boundaries fall, what gets a queue and what does not. An attribution platform is judged on numbers being right, so the cost of getting those choices wrong is not a slow page. It is a customer optimising against a lie.
Around both: leading internal and external teams, which in practice means writing things down clearly enough that people in different places and time zones arrive at the decision you would have made yourself.