The Agency's Midyear Website Audit: Which Visits Became Real Conversations?

A midyear client review should answer a plain question: which website visits became conversations the client wanted to have?

A chart of rising traffic cannot answer it. Neither can a list of contacts sent to a CRM. The agency needs a path from the campaign to the relationship, with enough detail to distinguish a useful conversation from a form submission, a duplicate or an unanswered inquiry.

Use the review to establish what you can prove, find the gaps and agree on the next decision. The result should be a working document the account manager and client can use together.

Agree on what the client bought

Start with the outcome in the engagement, not the largest number in the reporting tool.

If the agency was hired to generate suitable sales conversations, write down what makes a conversation suitable. That might mean a business in the client's service area, with a relevant need and a completed discussion with the right team. If the engagement was designed to support online purchases, use the purchase outcome instead.

Do not retrofit a success definition to the available report. If the original goal was unclear, say so and agree on a usable definition for the next period. Keep the earlier period's limitation visible.

For agencies using visitor information to demonstrate value, this is the connection worth making: what the agency brought in, what the client received and what happened after the handoff.

Choose the group before counting the outcomes

A midyear audit covers a calendar period, but not every lead in that period has had the same time to develop.

Someone who first inquired in January has had months to reach a sales decision. Someone who inquired on June 29 has had very little time. Combining both without context can make recent campaigns look weaker for the wrong reason.

Group contacts by an agreed starting event, such as the week of their first eligible inquiry. Record the cutoff used for the audit. Then follow that group into the client's sales process.

Keep two useful views. One shows what happened during January through June. The other follows the people who entered during a particular campaign or month. The first helps with operating activity; the second helps you understand how those contacts developed.

Give every number a label that makes the view clear.

Build the account manager's worksheet

You do not need a new dashboard to begin. A shared worksheet can hold the evidence and the unresolved questions.

Create one entry for each campaign group, with these fields:

• Goal and qualification rule: the outcome the client agreed mattered.
• Starting event and dates: who belongs in the group and when they entered.
• Campaign and destination: the offer, source and page people actually saw.
• Eligible contacts: records left after the agreed duplicate and test checks.
• Sales disposition: accepted, rejected, waiting for review or missing.
• Completed conversations: what happened, with the client's definition attached.
• Opportunities: accepted sales opportunities associated with the group.
• Evidence gap: the missing connection or decision preventing a firmer conclusion.
• Owner and next action: who will resolve the gap and what they will check.

Store references to the underlying records where the review team has appropriate access. Do not copy unnecessary customer details into a presentation that will circulate more widely.

Reconcile a small sample together

Before debating totals, trace a few records with the client.

Choose one contact that progressed, one that was rejected, one still waiting and one whose source is unclear. Follow each from the campaign through the website action and into the relationship record.

Check the identity association, timestamps, owner and outcome. Ask whether an existing contact was updated or a new record was created. Confirm what “accepted” means in the client's process.

Review the integrations supporting the handoff against those examples. A connected system still needs to deliver the context the receiving person uses.

If the sample exposes a recurring issue, identify which totals it could affect. A few inspected records can reveal a broken rule, but they do not automatically establish how common the problem is across the whole campaign.

Show the difference between received and accepted

Consider a hypothetical client campaign with made-up results. During the quarter, the website recorded 120 successful inquiries. The review removes 10 repeated submissions and five internal tests, leaving 105 eligible inquiries.

The client reviews 90 of them. Sixty are accepted as suitable for a sales conversation, and 30 are rejected with recorded reasons. Fifteen still have no review decision.

Of the 60 accepted inquiries, 36 have completed a conversation by the cutoff, and 12 have become opportunities. The remaining accepted contacts may still be scheduling, waiting or no longer progressing.

That is a more useful account than “120 leads delivered.” It shows a website result, a quality decision, a follow-up outcome and a gap.

It also gives the account manager specific work. Find out why 15 inquiries were never reviewed. Check the reasons for rejection. Understand what prevented some accepted inquiries from reaching a conversation. Do not assign all three problems to campaign targeting.

Name missing evidence without assigning blame

An agency can control the campaign while the client controls the sales response. Both still need a shared account of the result.

Use neutral descriptions. “No sales disposition is recorded for 15 inquiries” is observable. “The sales team wasted 15 leads” claims an explanation the record may not support.

The same rule applies in the other direction. A client saying “these leads were bad” needs a definition and examples before the agency can change targeting usefully.

Separate missing information from a negative outcome. An unreviewed record is not a confirmed rejection. A meeting that has not happened by June 30 is not necessarily a permanently lost opportunity.

Give every unresolved issue an owner. Shared responsibility becomes practical when the next action belongs to a specific person.

Account for changes in collection

Before comparing January with June, check whether the site and reporting process stayed comparable.

Record changes to landing pages, forms, campaign naming, duplicate rules, sales stages and the tools observing the visit. A new collection setup can change the report even when the audience behaves similarly.

Include consent and visitor choices in that review. The agency should understand which observations were available in each period and where the report has limits.

Do not fill an unknown-source group with a favored channel to make the chart complete. Preserve the unknown category and investigate whether the connection can be established.

Leave the meeting with one consequential decision

The audit should end with a decision supported by the evidence. That could be correcting a handoff, narrowing an offer, improving a page or changing the client's review process.

Write the decision, owner, review date and expected observation. For the hypothetical campaign, the first action might be assigning the 15 unreviewed inquiries and agreeing on a consistent rejection reason. That could improve the next discussion before anyone changes a campaign budget.

Use the path from website activity to action as the thread through the meeting. Bring the client a result they can inspect, a limitation they can understand and a next step someone actually owns.