Most Meta Ads accounts that “do not work” are not failing at the ad. They are failing somewhere in the chain between the ad and the enquiry, and because the ad is the visible part, that is where the budget and the blame go.
Here is how lead generation on Facebook and Instagram actually works, and where it usually breaks.
The one difference that explains everything
Google Ads captures demand. Someone types “SEO agency in Delhi”, and you appear. The intent existed before your ad did.
Meta Ads creates demand. Nobody opens Instagram looking for you. Your ad interrupts something, so it has to earn attention first and only then ask for something. This single difference dictates everything else: your creative has to do the work a search query was doing, and your offer has to be small enough to accept from a standing start.
Running Meta like search, straight to a “Request a Quote” form, is the most common and most expensive mistake.
The five parts, in the order they fail
1. Creative
On Meta, creative is the targeting. The algorithm finds people who respond to what you have made, so the ad decides who sees it more than your audience settings do. Practical consequences: test genuinely different angles rather than colour variants, lead with the problem rather than the brand, and make the first second of video carry the message because most viewers never reach the second.
2. Offer
Cold audiences do not fill in long forms for a sales call. Something with obvious standalone value converts far better: a specific audit, a real teardown, a pricing guide, a checklist. The offer should be worth having even if the person never buys.
3. Destination
Two workable options, with a real trade-off. Instant forms inside Meta convert at a higher rate because nothing loads, but the leads are often softer. A landing page filters harder and gives you room to make the case, at the cost of some volume and a dependence on load speed. Sending paid traffic to your homepage is not a third option; it is a leak.
4. Tracking
This is where most accounts quietly fail. Without the Meta Pixel plus the Conversions API sending server-side events, Meta is optimising toward clicks rather than the outcome you care about. Browser-only tracking loses a meaningful share of events to privacy controls, so the algorithm gets a distorted picture and spends accordingly.
5. Follow-up
A Meta lead is warm for minutes, not days. The person was scrolling and did something small. Response time is often the highest-leverage variable in the entire funnel, and it is the one part that has nothing to do with the ad account.
A structure that holds up
| Stage | Audience | Job of the ad |
|---|---|---|
| Cold | Broad, interest or lookalike based | Earn attention, name the problem, offer something small |
| Warm | Video viewers, page and profile engagers | Build credibility, show how you work |
| Hot | Site visitors, form starters, past leads | Ask directly, remove the last objection |
Most accounts run only the cold layer and judge the channel on that. Retargeting is where the cheap conversions are, but it does not exist until the cold layer has built an audience to retarget.
Reading the numbers without fooling yourself
Cost per lead is the headline metric and the most misleading one, because it says nothing about whether the leads were real. Track cost per qualified lead, and cost per customer, even if you have to do it in a spreadsheet at first.
Two supporting checks worth watching: hook rate, the share of viewers still there after three seconds, which tells you whether the creative is the problem, and frequency, which tells you when an audience has seen it too often and results are decaying for reasons unrelated to the ad’s quality.
Why campaigns break after working
Almost always creative fatigue. The same ad to the same audience decays, frequency climbs, cost per result rises, and it looks like the platform changed. The fix is a creative pipeline rather than a creative. Plan to replace, not to optimise a winner forever.
The second most common cause is editing a campaign that was working. Significant changes reset learning, so performance drops, which invites another change. Build the next test as a new campaign instead.
Frequently asked questions
What budget do I need to start?
Enough to gather data before you judge it. If your target cost per lead is one thousand rupees, a few hundred rupees a day will take weeks to produce a readable result, and you will make decisions on noise. Better to run one properly funded test for a few weeks than three underfunded ones.
Meta Ads or Google Ads first?
If people already search for what you sell, Google first, because you are capturing existing intent. If your category is not searched, or you need to explain why the problem matters, Meta first. Most businesses eventually need both, because they do different jobs.
Should I use instant forms or a landing page?
Instant forms give more leads at lower quality; landing pages give fewer at higher quality. If your sales team has spare capacity, take the volume. If their time is the bottleneck, filter harder with a page and qualifying questions.
Why did my leads get worse after a good first month?
Usually creative fatigue combined with audience saturation. Check frequency: once the same people have seen the ad repeatedly, the responsive ones have already responded and you are paying to reach the rest. New creative, not a new budget, is the fix.
Do I really need the Conversions API?
If you care about lead quality, yes. Browser-only tracking loses events to privacy controls, so Meta optimises on partial data. Sending server-side events gives the algorithm a truer signal, and the difference usually shows up in lead quality rather than volume.
Where to look first
Before changing a single ad, check whether your conversion tracking is actually reporting qualified leads, and how fast a new lead gets contacted. Those two fix more campaigns than creative changes do.
If you want a second pair of eyes on an account, see how we run Meta Ads or tell us what is happening.