Layer 6 · Measurement
Source-to-customer attribution and reporting a founder can read, built at a scale a small business can actually maintain.
When the website, form, CRM, and revenue records do not share a source identifier, a channel-to-revenue answer becomes an estimate. That is a system-design problem rather than a reporting failure.
Analytics platforms measure sessions. Ad platforms measure their own clicks. The CRM measures deals. None of them, on its own, connects an early search or campaign to a later invoice.
Measurement makes other decisions easier to test. Without it, improvements rely more heavily on assumptions and budget conversations cannot use a shared definition of success.
Fit
Symptoms
Scope
The small set of numbers that would actually change a decision, defined and written down before any tag is placed. The work starts with the question so the implementation does not become a collection of unused events.
Meaningful actions tracked with a consistent naming convention: CTA engagement, form starts, form submissions, booked calls, and opt-ins. Named so that a report six months from now is still readable.
Channel, campaign, and landing page captured at first touch, persisted through the inquiry, and written into the CRM record so it survives to the closed deal.
The join between marketing activity and revenue: which channel produced qualified inquiries, which produced customers, and what those customers were worth. This is the answer the measurement layer exists to produce.
Median first-response time, qualification rate, booking rate, and workflow failures. These operational measures are assessed alongside marketing measures so lower-cost interventions are not overlooked.
Server-side capture where it is more reliable and less invasive, no unnecessary personal data, and disclosures that match what the site actually does. Tracking you would be comfortable explaining to the person being tracked.
An event dictionary and a definition of every metric, so the numbers mean the same thing after you hire someone or change tools.
Process
We list the decisions you actually make — increase spend, cut a channel, hire, change an offer — and work backwards to the smallest set of numbers that would inform them.
What is currently tracked, what is broken, what is double-counted, and what is quietly missing. Existing analytics get improved rather than replaced wherever that is the cheaper honest answer.
Events implemented, source captured at first touch, and the join to the CRM built so revenue can be traced back to its origin.
A view a founder reads in two minutes, an event dictionary, and a written definition of each metric so the numbers survive staff and tooling changes.
Outcomes
These are the kinds of change this work produces. No numbers appear here, because no client result has been measured and published yet. When one has, it will appear in Work with its baseline, method, and time window.
Tooling
Categories, not endorsements. The right answer is usually the smallest reliable set, not the most impressive one.
The connected view
A growth system is only as strong as its weakest handoff. Fixing one checkpoint in isolation helps only if the adjacent checkpoints can carry the result.
See the whole systemCheckpoint 9 feeds evidence back into checkpoint 1. That loop is the point.
Where this sits in the system
Straight answers
If yours is more specific, put it in the audit form.
support@ariadne.fyiFor traffic behaviour, often yes. For revenue attribution, no — GA4 does not know which deals closed or what they were worth. The value is in the join between web behaviour, the inquiry record, and the CRM outcome, and that join has to be built.
The implementation is designed to collect the minimum needed to answer the questions we agreed on, favours server-side capture, and comes with disclosures that match what the site actually does. If a measurement would require data you should not be holding, I will say so rather than build it.
Partially, and honestly. First touch and last touch are both recorded, so you can see the shape of the journey. Assigning a precise fractional value to each touch is a modelling exercise that produces confident-looking numbers with weak foundations, and I would rather show you both ends than invent the middle.
Usually not. The first step is auditing what exists and what it is quietly getting wrong. Fixing and extending an existing implementation is normally cheaper than replacing it, and losing historical data has a real cost.
Your practical starting point
The audit reviews what is tracked today, what is missing, and which unanswered question matters most, alongside the other diagnostic checkpoints.