The most expensive lead isn't the one with the highest CPL—it's the one nobody calls back

Imagine generating 1,000 leads at an average CPL of €30. You have invested €30,000, and the marketing report looks healthy: campaigns are active, forms are being completed and cost per contact is under control.

Then you open the CRM. One hundred and eighty leads have no recorded attempt. Others were called once. Some say “call back,” but have no date, owner or next activity.

Is the CPL still €30? Only on the advertising dashboard.

The company spent €30,000 to generate 1,000 contacts but actually worked only 820. Its effective cost per worked lead has already risen to €36.59, before quality, appointments or sales even enter the calculation.

This is the part of lead generation that rarely appears in reports: the problem is not only what it costs to generate a lead, but how much of that investment actually enters the sales process.

CPL measures acquisition, not handling capacity

Cost per lead answers one question: how much did we spend to acquire a contact?

It does not answer what happens next:

- How many leads were accepted and worked?
- How long did the first attempt take?
- How many received a second attempt?
- How many were qualified?
- How many appointments were actually booked?
- How many activities were never recorded?
- How many leads remain in a generic CRM stage?

A campaign can be efficient while the sales process loses a meaningful share of the value it creates. Marketing may celebrate a falling CPL while sales complains about poor leads. Both teams can be right: contacts were acquired at a competitive cost, but some were handled too late, never worked or abandoned after too few attempts.

The break occurs between acquisition and ownership.

The problem is old, but it keeps repeating

In 2011, James Oldroyd, Kristina McElheran and David Elkington studied how 2,241 US companies handled online leads. According to Harvard Business Review, 23% did not respond at all; among companies that responded within 30 days, the average response time was about 42 hours.

This is an older US study, not a universal benchmark for every market in 2026. Its continued relevance is narrower and more useful: leakage does not necessarily come from missing tools. It often comes from unclear operational ownership.

Ten years later, the InsideSales Lead Response Study 2021 analysed more than 55 million sales activities across 5.7 million inbound leads and over 400 companies. It found that 57.1% of first call attempts happened after more than a week and that, within the intervals studied, conversions were more than eight times higher when the first attempt occurred in the first five minutes.

Five minutes is not a magic threshold. A lead does not suddenly expire in minute six. But delay increases the likelihood that the person speaks to a competitor, forgets the context, changes priorities or interprets silence as disorganisation.

Speed does not guarantee a sale. Delay does reduce the number of opportunities that reach sales in their best condition.

The missing metric: Lead Coverage Rate

CPL needs an operational companion:

Lead Coverage Rate = leads with at least one valid activity ÷ leads acquired × 100

A valid activity is not a generic CRM note. It is an attempt that was completed, dated, assigned to a person or system, given an outcome and connected to a next action when one is required.

If you generate 1,000 leads and only 820 have a valid activity:

820 ÷ 1,000 × 100 = 82% coverage

The remaining 18% is not yet a conversion problem. It is an execution problem. You cannot conclude that those leads were poor quality because the process never gave them a fair opportunity to be assessed.

Do not confuse unworked cost with potential margin

Two separate calculations are needed to estimate the impact of untouched leads.

1. Unworked acquisition budget

Unworked leads × CPL

With 180 unworked leads and a €30 CPL:

180 × €30 = €5,400

This does not automatically mean €5,400 has been lost. The contacts still exist and some may be recoverable. It means that this portion of acquisition spend has not yet had access to a complete sales process.

2. Unrealised potential margin

Use this formula:

Unworked leads × appointment rate × show rate × close rate × average margin

The rates should come from your company's actually worked leads, ideally segmented by source and time period.

Assume the historical data shows a 12% appointment rate, 70% show rate, 20% close rate and €2,000 average margin per customer:

180 × 12% × 70% × 20% × €2,000 = €6,048

That gives €5,400 in acquisition spend that has not been worked and €6,048 in estimated unrealised margin. Do not add them together as if they were the same measure. The first is a historical cost; the second is a probabilistic estimate.

The bottleneck may not be budget

If 250 leads arrive each week and the team can genuinely handle 180, the company adds 70 contacts to its backlog. After four weeks it has generated 1,000 leads, had capacity for 720 and accumulated 280.

Increasing advertising spend by 20% does not solve the problem. It makes it larger.

The right question is not “how many leads can we generate?” It is how many leads can we accept, qualify and move to the next step within the promised timeframe?

That is Lead Handling Capacity:

available operators × leads handled per operator × operating days

The calculation must include the entire workload: first attempt, later attempts, CRM updates, qualification, confirmations, calendar booking, rescheduling and handoff. Counting only first calls materially understates the load.

“We called them” is not enough data

A number that does not answer the first call is not automatically a lost lead. The person may be in a meeting, driving, working, using silent mode or simply unavailable at that moment.

The number of leads called should therefore be accompanied by at least:

- average attempts per lead;
- attempt distribution by time of day;
- contact rate;
- meaningful conversation rate;
- exclusion reasons and opt-outs;
- appointments booked;
- average time between attempts.

Indiscriminately increasing call volume is not a strategy. A sensible sequence defines frequency, timing, context and a clear stop when someone refuses further contact.

Before any campaign, verify data provenance, lawful basis, recorded preferences and applicable do-not-call registers. In Italy, the Garante per la protezione dei dati personali explains the right to object to marketing and the effect of the Public Register of Oppositions on earlier consent, subject to specific exceptions. This article is not a substitute for legal advice on a particular campaign.

A real example: 1,548 attempts are not 1,548 leads

In a 15-day Agendly campaign in the solar industry, covering both new and previously collected contacts, the recorded results were:

- 1,548 calls made;
- 207 meaningful conversations;
- 23 concrete leads generated;
- follow-ups, summaries and confirmations handled automatically.

Two useful indicators follow:

Meaningful conversations per attempt: 207 ÷ 1,548 = 13.4%

Leads generated per meaningful conversation: 23 ÷ 207 = 11.1%

The point is not that every call generates a lead; that would be false. The point is that producing 23 opportunities required volume, consistency, repeated attempts, qualification and reliable outcome tracking.

A salesperson did not need to manage all 1,548 attempts personally. They could focus on the part of the process where human expertise has the greatest impact. These results belong to this specific campaign and do not predict the performance of other databases.

A salesperson should not be the automation layer

The Salesforce State of Sales 2026 reports that the average salesperson spends about 40% of their time selling. The same study says 42% of sales reps feel overwhelmed by too many tools.

If a salesperson must continuously check new forms, call every lead, retry at different times, update every field, send confirmations, book meetings, remember every follow-up and still sell, the company is not protecting human relationships. It is using a skilled person as middleware between disconnected systems.

Automation makes sense when it absorbs repetitive work and gives people a contextualised conversation. It does not make sense when it tries to replace advice, empathy, negotiation or technical expertise.

The CRM audit to run today

Take the last 90 days and extract the following for every lead:

1. lead ID and source;
2. acquisition date and time;
3. first-attempt date and time;
4. response time;
5. number of attempts;
6. outcome and exclusion reason;
7. qualified yes/no;
8. appointment booked and attended;
9. sale, value or margin;
10. owner and next scheduled activity.

Then calculate:

Lead Coverage Rate

Leads with at least one valid activity ÷ leads acquired. The first objective is to locate leakage, not to choose an arbitrary benchmark.

Median response time

Use the median, not only the average. A small number of leads answered after several weeks can distort the average and hide normal behaviour.

Backlog Rate

Leads still without an outcome after the agreed maximum time ÷ leads acquired.

Contact Rate

Leads with a real conversation ÷ leads worked. Measure it by source, time of day and attempt number.

Qualification, Appointment and Show Rate

Qualified leads ÷ real conversations; appointments booked ÷ qualified leads; appointments attended ÷ appointments booked.

Recovery Rate

Reactivated leads ÷ leads added to the recovery campaign.

This analysis reveals whether the problem sits in campaign quality, response time, capacity, follow-up, qualification, calendar management or closing. Until those stages are separated, “the leads do not work” remains an opinion.

Three practical targets for the next 30 days

1. Move coverage close to 100%

This does not mean getting an answer from everyone. It means making sure every valid lead receives the intended action and every outcome is recorded.

2. Eliminate leads without a next activity

Every open contact needs an owner, a date, a next action or a closure reason. “Call back” is not a sufficient status.

3. Separate first contact from selling

The first layer should verify reachability, interest, requirements, timing and appointment availability. Sales should receive enough context to start a useful conversation, not another list to clean.

Where Agendly fits

Agendly can connect to the acquisition flow to:

- contact new leads quickly;
- work previously acquired databases;
- execute scheduled attempts;
- collect qualification information;
- record outcomes;
- manage calls and WhatsApp within the defined process;
- book appointments in a shared calendar;
- send confirmations and reminders;
- hand consultation-ready contacts to the team.

The value is not simply making more calls. It is making it verifiable that every acquired lead entered a process.

The campaign to optimise begins after the form

When CPL rises, marketing takes action. When close rate falls, sales leadership takes action. But when a lead sits between those two teams, nobody may act at all.

That is where budget becomes backlog. That is where an operational problem is mistaken for poor quality. That is where a company keeps buying contacts while it is still losing the ones it already has.

Before asking for more leads, measure how many contacts already in the CRM received a timely attempt, consistent follow-up, qualification, an outcome and a next action.

You may discover that the campaign you need to optimise is not the advertising campaign. It is the one that begins one second after the form is submitted.

FAQ