An advertising report can look convincing: ads receive impressions and clicks, visitors submit forms, and the cost per enquiry stays within plan. Yet these indicators do not answer the business owner’s key question: which advertising generated a qualified enquiry, a completed sale, and a confirmed payment? When analytics ends with the form while customer work remains in a CRM, ERP, cash register, or sales spreadsheet, the connection between spend and revenue can only be assessed indirectly.
Measuring digital marketing through sales requires an uninterrupted data route. The source of the click-through is retained on the website, passed to the CRM with the enquiry, supplemented with the qualification result, and matched with the deal outcome. The information then returns to a management report. This helps distinguish campaigns that create a flow of leads from campaigns that bring suitable enquiries. Even a well-configured connection does not make attribution an absolute truth: some contacts will inevitably remain unknown.
Why impressions, clicks, and cost per lead are not sales
Impressions describe ad delivery, clicks describe click-throughs, CTR is the ratio of clicks to impressions, and cost per lead is the cost of a recorded enquiry. These metrics are needed to diagnose ads and landing pages. The mistake begins when an intermediate action is treated as the final business outcome. A submitted form can be a duplicate, an irrelevant request, an enquiry from an unsuitable region, or a contact the team could not reach.
Two campaigns may generate the same number of leads but different numbers of qualified enquiries and sales. A cheaper lead is not necessarily more valuable: without a status and processing outcome, it is only a record that an action occurred. In a B2B process, each enquiry should therefore have a source, status, and result. Marketing can then be evaluated not only by the number of entries into the funnel, but also by what happened after the contact was handed to sales.
This does not mean every metric should be replaced with one revenue figure. Clicks help check ads, website events help check the user journey, and enquiries show the volume of incoming demand. A budget decision needs the next level: qualified enquiries, completed sales, and confirmed revenue assigned to a channel under a rule chosen in advance.
The data chain from an ad to payment
It is more useful to view measurement architecture as a sequence of events than as a list of applications. A visitor follows an ad, reaches a landing page, takes a meaningful action, and submits an enquiry or requests a call. A CRM record is created, a manager qualifies the enquiry and records the outcome. If payment is accounted for in another system, the completed sale is additionally matched with a confirmed financial transaction.
- Advertising click-through. Available information about source, channel, and campaign is recorded.
- Landing page. The entry page and time of the known contact are retained.
- Conversion event. A form submission, call request, order, or another predefined action is registered.
- CRM record. Marketing fields and a secure internal identifier are transferred with the enquiry.
- Qualification. A manager records whether the enquiry meets company criteria and, if it does not, the reason for rejection.
- Sale and payment. The deal receives a final status and the amount is recorded under an agreed, verifiable rule.
- Feedback loop. The outcome is matched with the original campaign and included in the marketing report.
Technical integration alone is insufficient. Marketing needs to maintain campaign naming rules, the integration team needs to ensure field transfer, and sales needs to update statuses promptly. The report owner defines what counts as an enquiry, qualified enquiry, completed sale, and confirmed revenue. If departments use different definitions, automated exchange will only transfer inconsistent data faster.
Data that must be retained
An initial working setup does not need to collect everything available. More important is stable transfer of a small set of fields: source, channel, campaign, landing page, contact time, conversion event, internal enquiry identifier, qualification status, and sale outcome. URL parameters can carry source, channel, and campaign labels, but they are useful only where naming rules are consistent and the values are retained until the CRM record is created.
- Source and channel — where the known click-through came from and which acquisition group it belongs to.
- Campaign — a stable name or identifier for an advertising initiative.
- Landing page — the first known page in the route.
- Date and time — data used to check the sequence of events.
- Conversion event — an action with an unambiguous trigger condition.
- Internal identifier — the technical key connecting a website event to a CRM object.
- Status and outcome reason — qualified, rejected, sold, cancelled, or another approved outcome.
- Amount and currency — only from an agreed source that relates to the relevant sale.
Free text is poorly suited to joining data. If one employee records a channel as “social media”, another uses an abbreviation, and a third names the platform in a comment, the report will require continuous manual cleanup. It is more practical to use controlled lists of permitted values and store the technical code separately from the name that staff understand.
First source and contact before the enquiry
It is useful to distinguish the first known source from the last known contact before conversion. The first helps explain how a person entered the measurable funnel; the last shows after which contact the enquiry appeared. If a single source field is overwritten on every visit, the original history disappears. Rules for retaining and updating fields should therefore be defined before the report is built.
Campaign parameters describe a specific click-through, but they do not by themselves prove a person’s entire journey. When no permitted and sufficiently reliable identifier exists between events, records must not be joined merely because their timing or behaviour appears similar. It is better to label that part as unknown than to create an apparently precise link based on an assumption.
Where the advertising source is lost
One typical break occurs between the landing page and the form. The analytics layer records the campaign, while the form handler sends only a name, contact detail, and message to the CRM. Marketing then sees a conversion but sales receives an enquiry without a source. To prevent this, approved marketing fields need to travel automatically with the enquiry rather than be copied by an employee from a separate report.
The next control point is the transition between CRM objects. A source can remain on the lead record but disappear when a contact, opportunity, or order is created. Another risk is a duplicate that receives the current channel and replaces previously retained history. The entire record lifecycle needs checking. Where marketing data must move beyond the CRM and be matched with an order or payment, a designed CRM and ERP integration is required.
Redirects, multiple domains, external forms, telephony, booking systems, and payment pages require separate review. For each transition, a short technical protocol is useful: which fields were received, which key was created, where it was transferred, and which result was returned. A user journey that appears continuous does not mean that data followed the same route. The source can also be lost outside the code when managers do not close deals, choose arbitrary statuses, or record rejection reasons only in comments. Automation can transfer fields and require a value, but qualification criteria and discipline in working with statuses remain management responsibilities.
Why one attribution model does not provide the whole truth
Attribution is a rule for allocating an outcome among known touchpoints. A first-touch model highlights the source of initial acquisition. A last-touch model connects the outcome with the contact before conversion. A multi-touch approach distributes contribution among several events. These models answer different questions and can produce different conclusions from the same data set.
For example, a person first arrives through an ad, later returns directly to the website, receives an email, and then sends an enquiry. The first model highlights the ad, the last model highlights the email, and a distributed model includes several touchpoints. No option is universally correct. First formulate the management question, then select a calculation rule and apply it consistently to the periods being compared.
The report should state the model, the analysis period, and the share of records without a known source. Figures from an advertising account, web analytics, and CRM cannot be added mechanically: systems may count different entities, including events, visits, enquiries, or sales. If the attribution rule changes, the date of change must be recorded; otherwise, period comparisons become invalid.
Returning qualification and sales to the marketing report
Sending the source to the CRM solves only half the task. A return flow is also required: after an enquiry is processed, marketing should receive the qualification result, rejection reason, deal creation, sale outcome, and payment confirmation where the company process provides for it. Matching is performed using the internal identifier, not the client’s name or comment text.
Rejection reasons are best kept short and structured: irrelevant request, unsuitable geography, duplicate, incorrect contact, no current need, or another category that reflects the real process. One general status for every unsuccessful enquiry cannot distinguish an issue with the advertising audience from the loss of a potential sale at the next stage.
The return exchange can be organised through an API, webhook, intermediate database, or controlled import. The method depends on the existing systems, but the result needs to be reproducible. Before expanding the setup, verify a control sample: whether record counts match, whether duplicates arise, whether campaign data is retained through the sale, and whether the final status returns. Practical principles for this exchange are covered in the article on CRM and ERP integration without manual data transfer.
A minimum management report
The first management view does not require dozens of charts. It is enough to compare spend, enquiries, qualified enquiries, completed sales, and confirmed revenue by channel or campaign. Data quality should appear alongside them: records without a source, without a final status, and with a conflicting identifier.
- spend for the selected period;
- number of enquiries and cost per enquiry;
- number of qualified enquiries;
- number of completed sales;
- confirmed revenue in the agreed currency;
- share of records without a source or outcome;
- attribution model used;
- date of the last complete data update.
An amount from the CRM cannot automatically be treated as confirmed revenue if the deal has not yet been reconciled with payment. One date rule must also be selected: assign the outcome to the first contact, deal creation, sale closure, or payment. Each option answers its own question, but mixing several rules in one table makes conclusions unclear.
A practical scenario for a company in Serbia
If a company already has a website, forms, calls, CRM, and a separate place where payment is confirmed, work can begin by aligning precisely those systems. When calculations use dinars, currency, source, and sale outcome must have the same meaning in every part of the report. Enquiry channels and qualification rules should follow the company’s real process rather than a universal template.
VMTech views digital marketing as a connection between acquisition channels, the website, analytics, conversion, and CRM data, not as a report limited to impressions and clicks. The approach to designing such a system is described on the VMTech digital marketing page.
Privacy, consent, and data minimisation
Measurement should be designed with a limited data set. An internal technical identifier can link events, while management reporting can show statuses and aggregate indicators. Names, phone numbers, email addresses, correspondence, and managers’ comments do not need to be transferred to web analytics or the advertising layer merely to make matching easier.
Complete tracking across all devices, browsers, and channels cannot be promised. If there is no permitted technical connection, the source or part of the route remains unknown. The architecture must not rely on bypassing consent, browser restrictions, or data-processing rules. Specific legal grounds, fields, and settings need to be checked for the company’s systems and jurisdiction; this material is not legal advice.
Five implementation stages
- Inventory. List channels, landing pages, forms, calls, CRM objects, and places where payment is recorded.
- Data map. For every transition, define fields, internal keys, update rules, and possible errors.
- One-campaign pilot. Select one landing page and one conversion type. The aim is to test source transfer and outcome return, not to prove sales growth.
- Quality control. Reconcile record counts, duplicates, empty sources, statuses, and cancelled sales.
- Gradual expansion. After a stable pilot, connect new forms and channels while preserving shared reference lists and definitions.
When complex attribution is not needed yet
A company is not ready for a complex model if enquiries are registered inconsistently, managers do not update statuses, qualification criteria are undefined, or a sale cannot be matched with payment. In that situation, a multi-touch report creates an appearance of precision over incomplete data. A more verifiable first step is mandatory source, status, and outcome-reason fields, consistent campaign naming, and regular reconciliation of a small sample.
Before the next advertising budget decision, answer five questions. Is the source retained with the enquiry? Can an ordinary enquiry be distinguished from a qualified one? Does a sale have a final status? Is that status linked to confirmed payment? Does the report show the share of unknown and conflicting records? If the answers are unclear, begin with a data movement map and a one-campaign pilot.
Connecting the website, analytics, and CRM does not eliminate uncertainty completely. Its purpose is to separate confirmed links from assumptions, show where data is lost, and give marketing and sales the same definitions. Lead cost can then be evaluated in the proper context: alongside qualified enquiries, completed sales, confirmed revenue, and explicit limitations in data quality.
