01Check the definition before the targeting
The first thing to agree is what a lead is. A form submission is one thing. A person who asked to be contacted, with a real requirement, in your area, at a price you both understand, is another — and the second one is the only one that can become an order.
Campaigns inside platforms can be optimised towards submissions that exist only inside the platform: instant forms are quick to fill, so more people fill them, and the report looks successful while the phone stays quiet. That is not fraud or incompetence. It is a definition nobody pinned down before launch.
The number to argue about is qualified enquiries per hundred spent, not cost per lead. If close rate is decent, an expensive lead can be the cheaper customer.
02The follow-up is part of the campaign
Paid attention goes cold in hours. A submission from tonight is a different conversion rate in the morning, and a very different one next week. So before touching a single audience: who sees an enquiry arrive, from which device, what do they say first, and how long before they say it.
- A notification path that works. Mailbox spam folders are where campaigns go to look like failures. Every enquiry has to reach a person on a device they actually carry.
- A response time someone is accountable for. Not a policy — a name and a target.
- An opening line written by whoever ran the ad. "You filled a form" is a dead call. "You asked about the three BHK in Sector 76" is a conversation.
- One place the status is written down. A shared sheet with a timestamp and an outcome is enough at the start. Without it, nobody can tell you what the leads were worth.
When a business says the agency delivered junk, that sheet is usually the missing evidence on both sides.
03What the page does between the click and the form
Traffic arrives on a phone, often on mobile data, having read one headline. If the page does not repeat that promise, the visitor leaves before any form is seen — and the campaign pays for the exit.
- Message match. The headline on the page is the headline of the ad, in the same words.
- Proof near the form, not on a separate page: work, credentials, real photographs, reviews that can be checked.
- A price signal. A range is enough. People who cannot tell whether they can afford you do not enquire, and the ones who do may not be able to.
- Few fields. Every extra question is a person who does not finish — keep only what the follow-up actually uses.
- Two ways to reach you. Form, call, or message. In India a share of enquiries arrive as a WhatsApp message that no conversion tracking ever counted, so the reporting understates a working campaign.
This is a page problem before it is a traffic problem, and it is the one fix that improves paid and organic results at the same time.
04Then look at the offer
If the wrong people are enquiring, read the ad back. A discount attracts price-led buyers; a broad promise attracts the merely curious; a lead magnet attracts people who want the free thing and nothing else. The audience did not misbehave — it answered what was said.
And one honest possibility: the offer may be the problem rather than the marketing. If people enquire, hear the price and vanish, more reach sells the same objection harder. That is a conversation for the product and pricing side of the business, and no campaign budget fixes it.
05Read the numbers in the right order
- Response time, from submission to first contact.
- Qualified share, as reported by whoever makes the calls — not as reported by the ad platform.
- Cost per qualified enquiry, then cost per closed one.
- Close rate by source and by campaign, which is where the junk is usually hiding.
- Frequency and creative fatigue, which explain a slow drift over weeks.
- Only then: bids, budgets, audiences, and the settings everyone argues about first.
06Questions people ask us
Should we ask more questions on the form to filter people out?
One useful qualifying question, the one a human actually uses on the phone. Five extra fields buy a little quality and pay for it in volume.
Are cheap leads ever worth buying?
Yes, if they close. The number that matters is cost per customer. A ₹90 lead that closes one in ten is not worse than a ₹400 lead that closes one in eight — and at those ratios it is much better.
Do we need a CRM before fixing this?
You need a timestamp and an outcome for every enquiry. A shared sheet does that. A CRM is what you buy when the sheet stops being enough.
The agency says the leads are good. How do I check?
Ask for the list of submissions and call five yourself, today, in the order they arrived. That takes ten minutes and answers the argument.
Does this mean we should stop the campaign?
Not usually. It means the campaign and the thing after the campaign get reviewed together. Pausing spend while the follow-up is broken guarantees a quiet phone for a different reason.
07Where this fits
The service pages cover their halves of this: Meta Ads for considered purchases, Google Ads for what people already search, website development for the page they land on, and the way we work is the same four steps that would be applied to a campaign that looks busy and sells nothing. If the paperwork is the open question, the GST guide covers it; if the area is narrow, start with local services.