Measure it. Test it. Ship what works.
Analytics, experimentation and conversion optimisation, designed and implemented by the people who build the product, so tests ship quickly and winning changes ship properly.
GrowthConversion & analytics
Conversion work is usually slowed down by implementation. Tracking that doesn't match the funnel. Experiments waiting weeks for development time. Variants that flicker as the page loads. Winners that never make it out of the testing tool and into the real codebase. We design the measurement, build the experiment, and ship the result.
Often the biggest gains aren't clever tests at all: a faster page, a form with fewer fields, delivery costs shown earlier. We look for those first.
What we do
What we work on
- 01
Analytics instrumentation
An event model that matches your funnel, implemented cleanly and respecting consent.
- 02
Funnel and journey analysis
Where people drop out, on which devices, and why, from data rather than opinion.
- 03
A/B and multivariate testing
Experiments implemented server-side or at the edge where possible, so nothing flickers.
- 04
Landing-page systems
Components and templates so campaigns launch fast and stay on-brand.
Web development - 05
Ecommerce conversion
Product pages, basket and checkout, measured and improved.
Ecommerce - 06
Speed as conversion
Performance work prioritised by its effect on the funnel.
Technical SEO
Every test starts as a spec
Written before anything is built, so the result can't be argued about afterwards. This is the template, not a result.
- Observation
- What the data shows, and where in the funnel.
- Hypothesis
- If we change this, for these visitors, then this metric moves, because of this reason.
- Change
- Exactly what differs between control and variant.
- Primary metric
- One number that decides the test.
- Guardrails
- What must not get worse: speed, revenue per visitor, errors.
- Sample and duration
- Calculated up front, run in whole weeks, not stopped early.
- Decision rule
- What we'll do on a win, a loss and an inconclusive result.
Capabilities
Underneath
Measurement
- Event and conversion tracking plans
- Server-side and consent-aware analytics
- Dashboards people actually read
Experimentation
- Hypothesis and sample-size planning
- Server-side and edge experiments
- Guardrail metrics and clean shutdowns
Implementation
- Landing-page templates and component systems
- Form and checkout improvements
- Winning variants rebuilt properly, not left in the test tool
Approach
How it runs
- 01
Instrument
Make sure the numbers are right before trusting any of them.
- 02
Find the leak
Funnel analysis, session evidence and performance data point to where people drop.
- 03
Fix the obvious
Broken, slow or confusing things get fixed, not tested.
- 04
Test the uncertain
Experiments for changes where the answer genuinely isn't known.
- 05
Ship and clean up
Winners built properly into the product; losers removed completely.
Why Northcode
We don't just recommend experiments; we build them, in your codebase, and ship the winners properly. The same team can fix the slow page the analysis uncovers.
What we avoid
- Tests stopped the moment the graph looks good
- Tracking everything and understanding nothing
- Dark patterns that lift a number and cost you customers
Contact
Tell us where your software is today.
Whether it’s already live, half-built, fragile, or still an idea, we’ll give you an honest view of what it needs and what can wait.