In short

Judge a partner on how it works, not on the numbers it forecasts. Ask it to define the lead it will deliver, explain the full journey from first contact to sales handoff, show what its reports contain, state who owns accounts and data, and describe how it handles consent. A credible partner makes uncertainty visible before you spend; a weak one hides it behind volume targets.

Start with what you are buying

"Leads" can mean anything from a list of email addresses to a booked meeting with a decision-maker. Before comparing proposals, write down which event you want to pay for and which events you only want reported. Then ask every candidate to price and report against your definition, not theirs.

If a candidate cannot or will not adopt a written definition, comparisons between proposals are meaningless — each will be counting something different.

Examine the whole journey

Lead generation is a chain: audience, message, channel, landing page or call, information notice, capture, qualification and handoff. Strong advertising can be wasted by a confusing form; a good form is wasted if nobody calls back. Ask each candidate to walk you through every link and to say which ones it runs and which ones remain yours.

Treat forecasts as hypotheses

Platform estimates of reach and cost, and agency projections of leads per month, are planning tools. They depend on your offer, your market, your price and your follow-up. A professional partner will present them as assumptions, explain which ones are most uncertain and propose how to test them with a limited budget before scaling.

Be cautious with guaranteed volumes or fixed costs per lead offered before the partner has seen your data. They often lead to looser definitions later.

Clarify ownership and access

Advertising accounts, tracking configuration, landing pages, email lists and lead records should belong to you, with the partner working through access you grant. Ask what happens at the end of the contract: which assets are handed back, in what format and how quickly. This protects you from being locked in and makes switching realistic.

Look at a real report format

Ask for an anonymised or invented example report. A useful one shows spend, leads at each defined stage, rejected leads with reasons, what was changed during the period and what will be tested next. A report made only of impressions, clicks and form fills tells you about the channel, not about your pipeline.

Check fit with your sales team

The best partner for you is one your salespeople can work with. Involve the person who will receive the leads in the final meetings. Agree on response times, the handoff channel and how feedback flows back. A partner that asks about your sales process early is usually a partner that cares about outcomes.

Worked example

Example: comparing three proposals (hypothetical)

A consultancy receives three proposals. The first promises a fixed number of leads per month at a set price. The second offers a monthly fee and a detailed plan. The third proposes a two-month test with a small budget, a written lead definition and weekly reports including rejected leads.

The consultancy asks all three to price against its own definition: a request from a company with more than fifty staff, from an operations or finance role, with a project in the next year. The first candidate declines; the second agrees but keeps its own reporting format; the third adapts immediately. The consultancy chooses the test, keeping the account ownership and the right to stop after two months.

Questions to ask every candidate

  1. Which event do you price, and which do you only report?
  2. Which steps of the journey do you run, and which stay ours?
  3. Which assumptions in your forecast are least certain?
  4. Who owns accounts, tracking, pages and data, during and after?
  5. Which tools need consent, and how are opt-outs handled?
  6. Can we see a sample report with rejected leads and changes?
Choose the partner whose way of working you can inspect. Definitions, ownership and honest reporting predict results better than any forecast.

Examples in this guide are hypothetical and illustrate the method; they are not client cases.