Data you actually own
Self-hosted on infrastructure you control, not processed inside someone else’s advertising ecosystem. You decide how first-party data is collected, structured, retained and used — and you can take the raw data with you.
DB Analytics
Our own self-hosted analytics platform, built on Matomo and extended with what ecommerce brands actually need: every attribution model side by side, a direct WooCommerce integration, and custom modules built for your business.
It runs on infrastructure you control, and it answers commercial questions rather than producing another dashboard.
Analytics for modern businesses
DB Analytics is a customised, self-hosted build of Matomo, the open-source analytics platform. We didn’t rebuild analytics — we shaped the implementation, the integrations and the reporting around what our clients keep asking, and we extend it whenever a business needs something standard analytics doesn’t record.
Self-hosted on infrastructure you control, not processed inside someone else’s advertising ecosystem. You decide how first-party data is collected, structured, retained and used — and you can take the raw data with you.
Six attribution models — first and last interaction, last non-direct, linear, position based and time decay — reported side by side on the same orders, so you can see which channels start journeys and which ones merely finish them.
A direct WooCommerce integration takes orders and revenue from the store itself rather than from a tracked event that may or may not have fired. In practice we see 98–99% of orders and revenue attributed.
Because we run the platform ourselves, we can record things no standard analytics tool collects. Weather for a seasonal retailer, stock levels, pricing changes, promotions — whatever actually moves your numbers.
Matomo’s privacy controls, configured as part of a wider privacy-first analytics and consent setup rather than bolted on afterwards. We handle tracking, consent and reporting as one design, so they agree with each other.
Referrals from ChatGPT and other AI assistants are a real and fast-growing source of traffic. Recent platform updates identify them properly, so they show up as their own channel instead of being quietly folded into Direct.
Multi-model attribution
Run a campaign on Meta and Meta will tell you it worked. Run one on Google and Google will say the same. Each platform reports its own conversions using its own attribution model, and the totals rarely reconcile with what your store actually took.
DB Analytics runs six models over the same orders and shows them side by side. First click tells you the entry point. Last click tells you the closing channel. We make budget decisions on position based, because that is the model that also pays for what happened in between — the retargeting that brought someone back, which last click would quietly hand to Direct.
| Channel | First interaction | Last interaction | Position based |
|---|---|---|---|
| £60,000 | £50,70015.5% | £55,3507.8% | |
| Google Ads | £42,000 | £37,80010.0% | £39,9005.0% |
| Meta Ads | £12,000 | £7,20040.0% | £9,60020.0% |
| Klaviyo email | £6,000 | £5,7005.0% | £5,8502.5% |
| Direct entry | £18,900 | £28,10048.7% | £23,05022.0% |
| Search engines | £5,200 | £3,10040.4% | £4,10021.2% |
| Websites | £6,800 | £7,3007.4% | £7,0503.7% |
| AI assistants | £1,240 | £1,240no change | £1,240no change |
| Social networks | £400 | £2,100425.0% | £1,750337.5% |
Where the credit goes, first interaction to last
Same 90 days. Same orders. Same £92,540. Move from first interaction to last and £11,400 of credit changes hands — because a customer who found you through an ad last week typed your name in this week. Open the Campaigns row and the cost is specific: Meta loses 40% of its credit while email loses 5%. Report on last click alone and you will cut the campaign that created the demand.
Revenue credited to each channel under each model, with the change against first interaction. Illustrative figures modelled on a real 90-day export; the three Campaigns sources sum to the Campaigns row, and all three models total the same revenue, because attribution redistributes the same orders rather than counting new ones.
Three real shapes
A table tells you the models disagree. A journey tells you why — because you can see the touchpoint each model is choosing to ignore.
Google Ads conversion
Day 0
Google Ads
Searches a product term, lands on the category page, leaves.
Day 4
Meta Ads
Retargeted, browses two products, adds one to basket.
Day 11
Direct — buys
Types the URL in, completes the order.
Last click hands the whole £340 to Direct. The ad that found this customer, and the retargeting that brought them back, both show as nothing.
Meta Ads conversion
Day 0
Meta Ads
Discovers the brand in feed, reads one product page.
Day 3
Organic search
Looks the brand up by name, signs up for the newsletter.
Day 6
Klaviyo email — buys
Clicks the welcome offer and orders.
Email closed it — but Meta created the customer and put them on the list in the first place. Position based is the only one of the three that pays both.
Search engines conversion
Day 0
Organic search
Finds a guide article, reads it, looks at two products.
Day 2
Direct
Comes back to compare, does not buy.
Day 2
Google Ads, brand term — buys
Searches the brand name, clicks the ad above the organic result, orders.
The uncomfortable one. A brand ad took the last click from organic that had already done the work — worth knowing before anyone judges that campaign on last-click ROAS.
Beyond the standard reports
Traffic and revenue are the easy part. The work that moves a number is in how long people take to buy, where baskets break, and what the sessions that nearly converted have in common.
Cohorts
Cohort reporting groups visitors by the day they first arrived, then tracks the orders they place over the following days — 5, 30, 90 — with the revenue attached. Time to purchase stops being a guess.
Most reporting treats a visit that converts three weeks later as a failure. This shows it wasn’t, and lets you go back to the sessions that returned to see what brought them back. It is the clearest read available on whether retargeting is earning its budget.
Custom segments
Segment the full visitor logs by added to cart, reached checkout, abandoned basket and purchased, then read them against each other. Drop-off stops being one conversion-rate number and becomes a set of patterns.
That is where conversion work actually starts: not with a percentage, but with the specific thing a few hundred people did just before they left.
Log-level analysis
Because the raw data is yours, we can export a thousand complete visitor journeys at a time and run them through ChatGPT or Claude with the goals and purchase signals attached, for a full read of what is blocking conversion.
It surfaces the shape of the sessions that nearly bought and didn’t — the kind of pattern no aggregated report will ever show you, because aggregation is exactly what hides it.
Custom dashboards
Every client gets their own overview rather than a default screen: sales by channel, campaign or country, conversion rate and traffic by whichever metric matters to them, arranged the way they actually read it.
If the question changes, so does the dashboard. There is nobody to ask for permission.
Custom modules
Standard analytics records what standard analytics was built to record. Because we run the platform ourselves, we can add to it — so the data includes the things that actually explain a good week or a bad one.
Weather tracking
A client selling seasonal goods doesn’t need a report telling them sales fell in week 32. They need to know what the weather was doing. We built weather tracking directly into the platform, so every session records the conditions it started in — alongside the visitor logs, journeys and every standard metric.
Conversion rate can then be read against the weather rather than guessed at, which changes how a season gets planned, how stock gets bought and when campaigns go live.
WooCommerce integration
DB Analytics integrates directly with WooCommerce, so orders and revenue come from the store’s own records rather than from a purchase event that may have been blocked, missed or fired twice. Across the stores we run it on, we typically see 98–99% of orders and revenue attributed.
That matters most when the numbers are being argued over: when reporting reconciles with the store, the conversation moves on to what to do about it.
AI assistant tracking
People increasingly find products by asking an AI assistant. With the latest platform updates we identify that traffic properly, and we are seeing a marked increase in conversions attributed to ChatGPT and others.
In most analytics setups those visits land in Direct, where they are invisible — which means a channel that is genuinely growing looks like it doesn’t exist.
Alongside Google Analytics
Most of our clients run both, because each is better at something. But you are not locked into that, and you will not lose years of data by changing your mind.
Conversions into Google Ads
We can feed conversions and conversion IDs from DB Analytics straight back into Google Ads, so smart bidding optimises against orders your store actually recorded rather than against a purchase event that may have been blocked or missed.
Our honest view is that GA4 is still the better companion for Google Ads conversions in most setups, and we will tell you when that is the case. This is there for when it isn’t.
GA4 import
We can import your historic GA4 data into DB Analytics, so year-on-year comparisons keep working from day one rather than starting again from zero.
If you would rather not run two platforms at all, you don’t have to. You keep the history and leave the product — which is the point of owning your own data.
Getting set up
Moving to DB Analytics does not mean losing your history or turning anything off. It runs alongside what you already have, and we handle the work.
We stand up your own instance, deploy tracking across the site and configure it alongside your consent setup so the two agree. Google Analytics stays exactly where it is.
We connect WooCommerce directly, bring in your ad platforms and search data, then build any custom modules your business needs — and reconcile the revenue against your store until the numbers agree.
You get a commercial read of what sold, which channels started those journeys and what changed — ending in a recommendation rather than a chart export. We keep tuning as your mix and your questions change.
DB Analytics team
A combination of technical expertise and commercial insight, working together to help you understand your data and act on it with confidence.
FAQs
Clear answers to common questions about tracking, reporting and improving data accuracy.
A privacy-first analytics platform designed to simplify tracking, improve data accuracy and provide clear, usable insights.
It focuses on clarity and reliability, avoiding many of the gaps caused by cookie restrictions and complex configurations.
Compliance comes from how a tool is configured, not from the tool itself — so no platform can grant it. DB Analytics gives you a far better starting point: self-hosted, on infrastructure you control, with you as the data controller and nothing passed into an ad ecosystem. We set up IP anonymisation, your own retention periods and consent handling alongside your consent banner so the two agree. Your privacy policy and lawful basis stay yours, and are worth confirming with your own advisor.
Usually yes, and we configure the banner and the tracking together so they agree.
Yes. Dashboards and reports are customised so you only see the data that matters to your goals.
Absolutely. Reporting is designed to guide decisions and highlight opportunities for improvement.