The SEO report shows impressions and clicks going up. The website report shows page visits. The sales report shows deals closed this month. Three reports sit on the same desk, and none of them answers the owner’s question: how much money was lost on the way from a search query to a payment. The budget for visibility has been spent, and nothing connects that budget to revenue.
The commercial route has seven stages: search demand → visibility → click → landing page → desired action → lead handling → sale. This article covers that route; site ownership, access rights, and brand presence in AI answers are outside its scope and are covered in separate articles.
Every stage has its own data and its own owner, and a loss at any one of them changes the economics of the whole channel, even when the earlier stages look healthy. The purpose here is to show how to check the route stage by stage, what data that takes, and how to pick the next priority by economic impact.
The commercial result is shaped by the whole route
The customer journey runs through four areas of responsibility. Acquisition belongs to marketing: demand, visibility, the click. Interaction belongs to product and development: the page, the flow, the form, or the cart. Passing the lead on belongs to analytics and integrations: an event on the site becomes a record in the CRM. Handling belongs to sales: contact, qualification, deal.
Each department sees its own area and reports within it. Marketing counts clicks, development counts errors, sales counts deals. Every departmental report ends at the boundary of its responsibility, so the economics of the chain appear in no report at all, and it takes business analytics as a function outside any single department to assemble them.
The stages are connected by one identifier that travels the full path: traffic source, landing page, lead, and deal outcome are written into a single record. In practice this means the source and page parameters are passed from the site into the CRM together with the inquiry, and the deal status is returned to analytics. Setting up that link is part of our end-to-end analytics service. Here one rule is enough: when the identifier breaks at any stage, everything after it is guesswork.
The owner leaves this section with a route map: four areas and the data owner for each stage. Without that map the next steps turn into an argument between departments.
Search visibility is where the route starts
Demand has to match business priorities
Search demand is not uniform. Some queries lead to high-margin categories, some to products with no markup, some to regions the company does not ship to. The line “organic traffic is up” says nothing about which share of that growth landed on priority categories.
The check is simple and runs on Search Console data compared with category margins from the accounting system: the share of impressions and clicks for priority categories is set against their share of margin. If visibility grows in categories that bring, say, 5% of margin, the channel is working for the report while profit stays where it was. Seasonality and geography are checked the same way.
Branded and non-branded queries are separated, and so are commercial intent and informational interest. “Buy an 80,000 BTU boiler” and “how to choose a boiler” bring different people to different pages with different expectations. Merging them into one “traffic” figure hides the real shape of demand. There is one more entry point to the route — AI answers that cite the site. Business presence in AI search is covered separately, and a visit from such an answer then passes through the same stages of the map.
The landing page defines the quality of the click
A click brings money only when the landing page matches the intent of the query and lets the visitor take the next step. A category, a product page, a service page, a regional page, and an article all do different jobs. A click from a commercial query onto an article with no path into the catalog ends in reading and leaving.
The second condition is the accuracy of commercial data. Price, availability, delivery and payment terms, lead time, and the next-step button have to match reality at the moment of the click. A page with last year’s price or an out-of-stock product receives the click and loses it within seconds, and in the SEO report that loss looks like a successful visit.
The owner checks two facts: whether visibility leads to the business’s priority offering, and whether the next step can be taken on the landing page. Both are checked on a sample of the ten most visited landing pages.
The technical foundation carries the customer journey
Page availability
The technical state of a site reaches the customer through a few observable things: the page loads in an acceptable time on a phone, elements do not shift while it loads, price and contacts are visible without extra clicks, and redirects lead to the right page. Every failure has a price in interrupted journeys, and that price is measured in data: errors in the log, sessions that end in the first second, the conversion gap between devices. Page experience metrics (loading, responsiveness, and layout stability) and how they relate to user behavior are covered in our article on Core Web Vitals, and a technical site audit is part of our technical SEO service.
Navigation and search
In an online store the journey runs through categories, filters, internal search, product options, and inventory. A filter that returns zero results, or an internal search that serves products from another category, stops the journey before the cart. Internal search data is valuable on its own: queries with no results show demand the site received and left unanswered.
On a service site the journey is different: from the service page and terms to proof of expertise and the inquiry form. A gap between the service description and the proof, or a form with no path to it from the service page, produces the same effect.
Forms and checkout
The last step of the journey is the most expensive one, because the customer who reached it has already decided to act. Forms, cart, registration, payment methods, delivery cost, and confirmation of the action taken are checked separately and regularly. Every release can change how a form behaves, so the release history belongs in the diagnostic data set.
External research gives a sense of scale. According to Baymard Institute, as of September 2025 the average documented cart abandonment rate across 50 studies is 70.22%. That is an average across different markets, mostly the US and the EU, and it describes the industry as a whole. For a particular store what matters is its own rate by segment, and among the reasons for abandonment Baymard lists extra costs at checkout, slow delivery, distrust of the site when entering card details, forced registration, and a checkout process that takes too long.
Checks like these make up the list of what gets done before any conversation about growing traffic. Pushing more volume into a journey that does not complete means increasing the losses.
Conversion shows the quality of a completed journey
Inside a company the word “conversion” often means a single number: the share of visitors who placed an order. That number hides more than it shows. An inquiry, a call, an order, a payment, and a repeat purchase are separate goals with separate journeys. Each has its own rate, and an average across the site guarantees nothing.
The formula itself is simple: conversion equals completed desired actions divided by sessions or users for the period. It only means something with three qualifiers: which goal, which segment, which period. Mobile order conversion in the “boilers” category in August and desktop call conversion across the whole site for a year are two different numbers, and they cannot be compared with each other.
Analytics terminology needs the same precision. In Google Analytics 4, desired actions are called key events, while the term “conversion” in the Google interface refers to Google Ads and is not shown in standard GA4 reports. In this article “conversion” means the business metric, the share of completed journeys; when analytics settings are described, the term is “key event.”
Two checks decide whether the metric can be trusted. First, event accuracy. The “order” event has to fire once per order, after confirmation, and not fire on a page refresh. Second, quality of the result. Cancellations, returns, duplicates, and unqualified leads are subtracted from the gross figure. Put simply: a store with 2% conversion and 30% returns earns less than a store with 1.5% conversion and 5% returns. So the owner assesses the quality of a specific journey and segment; an average across the site is not enough to decide on.
Lead handling completes the measurement
Passing data to the CRM
An inquiry from the site brings the customer’s contact into the CRM. Along with the contact should come the traffic source, the landing page, the product or service, and the campaign, and the deal record should then gain a status and an outcome. Every field that fails to arrive breaks the link between the route and the money. A source without a page makes the page impossible to assess; a page without a deal outcome makes its economics impossible to assess. Planning that work is part of our website and CRM integration service.
Speed and quality of handling
From here the route passes into the hands of sales reps, and the data changes its nature. Time to first contact, lost inquiries, duplicates, changes of owner, and the lost reason are handling metrics recorded in the CRM. The lost reason deserves its own field with a fixed list of values: “too expensive,” “no answer,” “out of stock,” “unqualified lead.” Without that field, sales and marketing blame each other with no evidence.
The link to the sale
The link between an inquiry and revenue has limits of accuracy, and those limits are worth admitting. The attribution model decides which channel a deal is credited to, and GA4 has three of them: data-driven, last click across paid and organic channels, and last click across Google paid channels. Direct visits are credited only when the entire path to the key event consists of direct visits; re-attribution is possible for up to seven days after the event. Other analytics systems have models of their own. The accuracy of the link to revenue is bounded by the chosen model and by how complete the passed data is; absolute accuracy does not exist, and an article that promises it is misleading.
The line of responsibility here is clear: the site’s job is to deliver inquiries with complete data, and from that point quality is measured by CRM metrics.
The loss map for the owner
All stages of the route come together in one tool. Every cause in it is written as a hypothesis: it becomes a fact only after the check named in the last column.
| Stage | Signal | Data | Likely cause | Owner | Action | How to verify |
|---|---|---|---|---|---|---|
| Demand | Low share of priority categories in impressions | Search Console + category margins | The demand map does not match business priorities | Marketing / commercial | Revise the demand map | Compare the share of impressions with the share of margin by category |
| Visibility | Impressions high, clicks low | Search Console: CTR by page and query | The snippet or the page does not match the query | Marketing / SEO | Align the page with the query | Sample of 20 queries with the widest gap between impressions and clicks |
| Click | Click lands, session ends immediately | GA4: engagement rate, bounce rate, device | The page is slow or opens with an error | Development | Fix the confirmed technical cause | Error log, test on mobile, comparison across devices |
| Landing page | Click lands, the action never starts | GA4 + session recordings + UX test | The page misses the intent or key data is out of date | Product / development | Review the page flow | Sessions, test with 5 users, check price and stock |
| Form / cart | Started, never finished | Events + form error log | An error or an extra step in the flow | Development | Fix the confirmed cause | Error log for the period, retest after release |
| CRM | Source or product goes missing | CRM + integration log | Fields are not passed or get overwritten | Sales Ops / development | Restore the data transfer | Sample of 20 inquiries: reconcile fields site → CRM |
| Handling | Long time to first contact, duplicates | CRM: response time, statuses, owners | No routing rule or no deadline control | Sales | Set a standard and control it | Distribution of response time over a week by rep |
| Sale | High share of deals lost after the inquiry | Lost reasons in the CRM | Lead quality or handling quality | Sales + marketing | Separate lead quality from handling | Lost reasons by source and page segment |
A completed map becomes the basis of a shared backlog for SEO, development, analytics, and sales. Every row with a confirmed cause turns into a task with an owner, and rows with no data turn into a task to collect it.
Priority is set by economic impact
After the checks, a handful of confirmed problems remain in the map, and there are never enough resources for all of them at once. Five criteria set the order.
First, the volume of demand or pages involved. An error on ten product pages and an error across a whole category carry different weight. Second, the effect on the desired action and how often the error occurs. An error that stops checkout in one case out of five weighs more than a cosmetic defect. Third, the cost of the fix. Fourth, confidence in the cause, backed by data. Fifth, how long it takes for a signal to appear after the change. A form fix can show up in the data within a week; a change to category structure may take a quarter.
An illustrative case. A store has two confirmed problems: in a category that accounts for 12% of total margin, the filter returns an empty result for a third of combinations, and on the homepage a banner overlaps the header logo on older phones. The first is less visible, yet it affects a commercially significant category, stops the journey before the cart, and takes two days to fix. The second is visible to everyone, yet it touches 3% of sessions and stops nothing. Priority goes to the first.
Priority scoring puts the work in order; it stays an estimate and gives no exact profit forecast. The team picks the next step by economic impact and evidence, then checks the result in the data after the change ships.
SEO and development work in one cycle
A completed loss map is the entry point into a working cycle: data → cause → decision → development → verification → next priority. Analytics points to the stage where the journey breaks. The cause is checked in logs, sessions, and tests. The fix is designed as a change to the site, built by development, and verified on the same data that surfaced the problem. The map is then updated, and the next row becomes a task.
In the Inward Labs model technical SEO and development are a single workflow: analysis, delivery, and verification of the agreed scope belong to one team, so a decision that comes out of analytics reaches the code without losing meaning in a handover between contractors. The company receives a report on which row of the map was closed, what data confirms it, and which row comes next. For the cycle’s work to survive a change of team, the company has to keep control of the website as a business asset: rights, access, and documentation are covered in a separate article.
What the owner should do
- Assign a data owner to each of the eight rows of the loss map — the seven stages of the route, with lead handling split into data transfer and sales handling. Owner: head of the company. Timeframe: one week.
- Check that the fields “source, page, product, campaign” reach the CRM from the site, on a sample of twenty inquiries. Owner: Sales Ops together with development. Timeframe: two weeks.
- Add a “lost reason” field to the CRM with a fixed list of values, and separate losses into lead quality and handling quality. Owner: head of sales. Timeframe: one month.
- Collect the confirmed problems into a single backlog and order them by the five priority criteria, the way the Inward Labs approach to a shared SEO and development cycle sets out. Owners: marketing, development, analytics, and sales together. Timeframe: once the map is filled in.
CHECK THE PATH FROM SEARCH TO SALE
Inward Labs will identify the points of loss, verify the data, and set priorities for a shared SEO and development cycle.
