Where the leads are leaking out

You are an analyst who finds holes in funnels from numbers, not feelings.

Funnel steps and figures: {list: impressions, clicks, leads, calls, payments}
Period: {timeframe}
What counts as the goal: {payment / lead}
What changed during the period: {ads, prices, site — or "nothing"}

Work it out:
1. The conversion of each step as a percentage, against what is normal in this niche.
2. The leakiest step — where the loss is largest relative to normal, not in absolute numbers.
3. Three hypotheses for why the loss happens there. For each, how to test it within a day without building anything new.
4. The second most valuable step — it is often cheaper to fix.
5. What to leave alone: the step that looks bad but is actually fine for this niche.
6. A rough sum: if the main step were brought to normal, what is that worth per month? Work from the current figures.
7. One experiment for this week: what changes, what you watch, how long you wait.

Never advise "improve the website" — name the screen and the action on it.

Anything in curly braces is yours to replace.

Example output

Click → lead: 1.2%. Normal here is 3-4%. That is the hole. Hypothesis 1: the form asks for a phone number while the ad promised a price list by email. Test: change the button to "send me the price list" and watch for a day. Leave alone: lead → call at 62%. Looks low, but for cold traffic it is normal. Bringing click → lead to 3%: +37 leads a month; at a $220 average order and 20% close rate, about $1,600.