Marketing lead quality analysis: measuring ad leads through conversations

Short answer

Measuring the quality of ad-sourced leads only by lead count or cost per lead is misleading. The cheapest leads are often the ones that buy least. Conversations show quality earlier and more precisely: why the customer wrote (intent), whether they could be your customer (fit), which product they asked about and how the conversation ended.

A practical model combines four measures: channel and campaign × intent × fit × CRM outcome. The result is a simple quality indicator per campaign that lets marketing steer budget towards leads close to a sale, not towards lead volume.

Why cost per lead is not enough

The ad platform tells you what a lead cost. But a "lead" there is just someone who wrote or filled in a form. One campaign might bring 100 leads at 1 manat each, 90 of whom write "how much is it?" and leave without waiting for an answer. Another brings 20 leads at 5 manat, 8 of whom ask about a specific product and date. The second looks more expensive but may be cheaper in terms of sales.

The ad platform does not see conversation content. It does not know who wrote what. That information is only in your conversation base.

Measure 1: channel and campaign

Which channel each conversation came from is usually known automatically. The campaign is not always. There are three ways to link it: a special parameter or opening message in the ad link, a separate number or landing page per campaign, or the campaign the customer mentions in the conversation ("I saw your ad"). The last is the least reliable, but better than nothing.

Measure 2: intent

What does the customer want? A simple value list: "wants to buy", "comparing prices", "general information", "about an existing order", "job seeker / spam", "unclear". The last two often show the ad reached the wrong audience. How to set up an intent field is covered in customer intent analysis.

Measure 3: fit

Does the customer match the profile you serve? Fit criteria vary by business: do they live in the delivery area, does their budget reach the minimum order, in B2B are the company's size and need a match. This information is usually visible in the conversation: "do you deliver to Ganja?" or "I need just one".

• Fit: meets all the main criteria.
• Partial fit: one criterion is unclear or borderline.
• No fit: outside the service area, below the minimum volume, and so on.
• Unclear: the conversation has no information.

Measure 4: CRM outcome

Conversations show intent and fit but do not prove a sale. The lead's CRM status (meeting set, offer sent, sale, lost) should be combined with the conversation result. Without that link, the quality indicator stays only an early signal.

A simple quality indicator

For each campaign, calculate three figures and show them side by side:

1. 1Intent shareConversations with a "wants to buy" intent as a share of all conversations.
2. 2Fit shareThe share of "fit" and "partial fit" leads.
3. 3Outcome shareThe share of leads ending in a meeting, offer or sale (from the CRM).
4. 4Cost per quality leadCampaign spend / (number of leads with intent and fit).

The fourth figure is the most useful: it replaces the ad platform's "cost per lead" with your own definition of quality.

A weekly report for marketing

• By campaign: conversations, intent share, fit share, outcome share, cost per quality lead.
• The "unclear" share for each campaign — if high, conversations break off early.
• The most common "no fit" reason — a correction for targeting.
• Three evidence conversations per campaign: one good, one bad, one typical.

Illustrative example

This is an illustrative example and the numbers are invented. A car parts store runs two campaigns. Campaign A (broad audience) brings 400 conversations a month, with a 15% intent share and 50% fit share. Campaign B (owners of a specific make) brings 120 conversations, with 45% intent and 85% fit. A's cost per lead is half of B's, but B's cost per quality lead is about three times lower.

Decision: part of the budget moves to B, and A is not stopped but its targeting is narrowed based on the "no fit" reasons (delivery area). The effect is checked two weeks later.

Limits

• Some customers come to the store after the conversation and buy, which the CRM does not link to the campaign.
• In small campaigns percentages swing randomly — assess several weeks together.
• Intent and fit are AI estimates and must be checked with a sample.
• Lead quality also depends on the sales team — a late reply to the same lead makes quality look lower.

A shared language for sales and marketing

Sales often says "marketing brings bad leads", and marketing replies "sales doesn't close them". A conversation-based quality indicator turns the argument into numbers: intent and fit are mostly marketing's responsibility, the outcome share mostly sales'. When both teams look at the same table, they talk about a specific campaign and a specific stage instead of about each other.

Lead quality in Vexvon

In Vexvon a conversation's channel is known automatically, while fields such as intent and fit are built by the company and AI picks from the customer's words in each conversation. Vexvon CRM stores the lead's status, channel, priority and loss reason, and you can ask the panel's AI assistant for lead statistics, their trend and a "channel × intent" cross-tab. Ad platform spend does not flow into Vexvon automatically — to compute cost per quality lead it has to be added separately. More: analytics and CRM.

Next step

Pick 40 conversations from the last two weeks of the two campaigns with the biggest budgets and mark intent and fit by hand for each. If a difference shows, automate the fields. We can set this up with your own campaigns in a demo.