Guide · Meta diagnosis
“Not converting” is three different problems: no leads, leads that don’t qualify, and qualified leads that haven’t closed yet. Split the symptom before you change anything, because each one has different causes, and the third often calls for waiting.
Short answer
Meta lead ads that don’t convert usually have one of three problems. No leads points to tracking, delivery, offer or form friction. Leads that don’t qualify point to form type, audience, creative promise or the optimisation event. Qualified leads that haven’t closed usually need sales follow-up checks and patience, not campaign edits.
Start here
Pull the last 30 days of leads with their status from your CRM or lead sheet. The numbers put you in one row.
| Symptom | What you see | Check first |
|---|---|---|
| A. No leads or too few | Spend and clicks, but few or no form submissions recorded | Tracking, learning status, recent edits, offer and form friction, delivery limits |
| B. Leads, few qualified | Healthy lead count; sales says most are junk or unreachable | Form type and questions, optimisation event, creative promise, placements |
| C. Qualified leads, no sales yet | Sales accepts leads, but no deals have closed in the period you are judging | Sales cycle length, follow-up speed, stage progression |
Short answer
Check four things in order: that the lead event or form is recording submissions, whether the ad set is learning limited or was recently edited, whether the offer and form ask too much, and whether targeting or a special ad category is restricting delivery. Change one thing at a time and give each change a week.
Short answer
Usually because Meta is optimising for form submissions, not for leads your sales team can qualify. Check the form type and questions, whether the ad promises something the offer can’t deliver, and whether sales follows up quickly. Then compare cost per qualified lead by ad and ad set, not cost per lead.
If Google and LinkedIn leads show the same pattern, the cause is probably shared, such as a vague offer or slow follow-up; see junk leads from ads.
If sales is accepting leads and nothing has closed, compare the period you are judging with your normal cycle. A consultancy whose proposals take 60–90 days to close can’t judge last month’s leads on revenue. Check instead:
Speeding up that loop is covered in the sales feedback loop guide.
Original framework
Each branch ends in one of four decisions: act, review, investigate or wait.
| If | And | Decision | Next step |
|---|---|---|---|
| A. No leads | Events or CRM counts don’t match submissions | Investigate | Fix tracking before touching campaigns |
| A. No leads | Several significant edits in the last week | Wait | Freeze edits for a week, then reassess |
| A. No leads | Tracking works, no recent edits, high-commitment offer | Act | Test a lower-commitment offer |
| B. Few qualified | More volume form, no qualifying question | Act | Switch to Higher intent and add one qualifying question |
| B. Few qualified | Quality differs sharply by ad or placement | Review | Pause the weakest after a fair number of leads |
| B. Few qualified | 200+ Instant Form leads a month, CRM stages available | Review | Assess Conversion Leads eligibility |
| C. No sales yet | Judging period shorter than the sales cycle | Wait | Judge on stage progression until the cycle completes |
| C. No sales yet | Leads contacted late or not at all | Investigate | Check sales follow-up, not the ads |
Worked numbers
₹90,000 a month buys 300 Instant Form leads at ₹300 each, but only 18 qualify: ₹5,000 per qualified lead. The form is More volume with no qualifying question. The fix is a Higher intent form with one question on order size. If the next month brings 140 leads and 21 qualified, cost per lead rises to about ₹643 while cost per qualified lead falls to about ₹4,286.
25 website leads a month, 10 qualified, and no proposals signed within 30 days. The firm’s cycle is 60–90 days. Campaigns stay as they are; the firm checks that every qualified lead was called within a day and reviews again when the first cohort completes a cycle.
Leads stopped after a week in which the team changed audiences twice and swapped creative three times. Tracking checks out. Every change was a significant edit, so the ad sets never left learning. The team stops editing for a week before deciding anything.
Short answer
Review weekly, change rarely. Each week, check delivery, frequency, creative fatigue and lead quality by ad. Each month, reallocate budget and retire angles based on cost per qualified lead. Avoid daily edits: significant edits can restart learning, and lead quality often takes weeks to show in your CRM.
| When | Check | Usual decision |
|---|---|---|
| Weekly | Spend pacing, delivery status, learning status, tracking health | Investigate anything broken; otherwise wait |
| Weekly | Qualified rate and speed to first contact, by ad | Review ads with many leads and few qualified |
| Weekly | Frequency and click-through trend at creative level | Review fatigued creative; batch replacements |
| Monthly | Cost per qualified lead by campaign and ad set | Act: move budget, retire weak angles |
| Monthly | Lead-to-meeting or lead-to-opportunity rate | Review the form and offer |
| Quarterly | Whether qualified leads became opportunities and wins | Review the qualification definition and optimisation event |
Review at creative level, not only campaign level. A campaign average can hide one ad producing most of the junk.
Short answer
Wait when an ad set is still learning after a recent significant edit, when too few leads have had time to be qualified, or when your sales cycle is longer than the period you are judging. Acting on small, incomplete numbers usually adds volatility without improving results, and resets the learning you have already paid for.
Zephra monitors Meta campaigns for spend anomalies, click-through and conversion rate drops and tracking health, uses lead quality from CRM data or a spreadsheet, and labels each finding act, review, investigate or wait. Changes follow your permissions and approvals, with learning-phase protection. See Meta Ads for B2B and how it works, or read lead quality for the definitions to start from.
Questions, answered
Mostly they are real people who submitted with little intent, often through a prefilled form. A Higher intent form and a qualifying question cut these. Store Meta’s lead ID so you can trace odd leads back to their ad.
Break qualified rate down by ad within the same ad set. If quality varies a lot between ads with the same audience, the creative is the stronger lever.
Test it when quality is the problem and volume isn’t. Website leads usually cost more and qualify more often; compare the two on cost per qualified lead.
At least until the ad set has left learning, or a week without edits if it is learning limited, plus enough time for sales to qualify the leads. For many B2B accounts that means three to four weeks.
Only if you send CRM stages back through it. Sending lead events alone improves measurement, not the kind of leads Meta finds.
Before you change anything
The free ad account assessment reviews wasted spend, campaign structure, performance constraints and measurement gaps. No CRM required, and no campaign changes.
