Usually it goes downstream of the ad account. On a $10,000 month at $4 a click, 2,500 clicks become 100 leads, 20 actual conversations, 8 booked calls, 4 held calls and 1 closed deal. Eighty of the 100 leads you paid $100 each for were never spoken to. Pausing spend does not recover them.
- The arithmetic: $10,000 spend, 1 deal, $10,000 cost per closed deal, 0.4x return on ad spend at a $4,000 deal.
- The biggest single lever there is contact rate — 20% to 45% is worth more extra revenue than halving your cost per click.
- Before you change anything: run the 30-lead callback test below — three hours, costs nothing, and it separates a traffic problem from a follow-up problem.
- Two quiet weeks is noise. Below 10 conversions a month, nothing shorter than six weeks is a trend.
- Pausing is right in one case only: the leads are real and reachable, but they are the wrong people.
Where the money goes when I’m spending money on ads but I’m not getting a return
A funnel does not leak in one place. It leaks at every join, and the joins multiply: 2,500 clicks at $4, a landing page at 4%, a 20% contact rate, 40% booking, 50% show and a 25% close rate compound down to one deal. Now hold the ad account still and change one stage at a time. Each row applies a single realistic improvement to that baseline, and nothing else.
| Stage | Baseline | Achievable | Closed deals from the same $10,000 | Extra revenue at a $4,000 deal |
|---|---|---|---|---|
| Baseline, nothing changed | — | — | 1.00 | — |
| Cost per click | $4.00 | $2.00 | 2.00 | +$4,000 |
| Landing page to lead | 4% | 6% | 1.50 | +$2,000 |
| Lead to conversation (contact rate) | 20% | 45% | 2.25 | +$5,000 |
| Conversation to booked call | 40% | 50% | 1.25 | +$1,000 |
| Booked to held (show rate) | 50% | 65% | 1.30 | +$1,200 |
| Held call to closed deal | 25% | 30% | 1.20 | +$800 |
Read the fourth column and the argument with your agency changes shape. “The ads aren’t profitable” and “I’m wasting money on Facebook ads” are usually the same complaint pointed at the wrong stage. Halving your cost per click — the thing you are about to demand — is the second-best move on the list and by far the hardest. Answering the phone more often is the best one, and it is free.
Two caveats. These do not stack — multiplying all six gives 13 deals, which will not happen, because the stages overlap. And the last column moves with your deal value, so substitute your own. Every stage and its governing metric is listed in the 17 measurable sales pipeline stages and what each costs.
How it works
The 30-lead callback test
Pull your last 30 leads
Thirty leads, not thirty days. Record when each enquiry landed and when a human first tried to contact them.
Ring all 30 yourself
Not your setter, not the agency. Two attempts each, at different times of day, inside 24 hours.
Score answered, recall, fit
Did they answer? Do they remember enquiring? Do they have the problem you solve and the means to pay?
Fix one stage, keep spending
The pattern across the three scores names the leaking stage. Repair that stage before you touch the budget.
MAKE MORE SALES.
Pay-Per-Result pricing — We scale sales HARD aligned to your interests, better than anyone else.
Is it as urgent as it feels? The noise floor
Probably less urgent than it feels. If your normal run rate is one deal a month, a fortnight with zero deals happens roughly 61% of the time even when nothing has changed — a Poisson distribution with a mean of 0.5. You cannot detect a broken campaign in a window containing less than one expected conversion.
The noise floor rule: below 10 conversions a month, no window shorter than six weeks is evidence of anything. Two quiet weeks is noise. Six is a trend. Six with a known change in them — new creative, a new landing page, a setter who left — is a cause.
Google Ads says the same thing in its own language: it “can take up to around 50 conversion events or 3 conversion cycles for the bid strategy to calibrate”, and a campaign re-enters learning on a “Setting change” or a “Composition change” — so every anxious edit restarts the clock you are complaining about (Google Ads Help, Duration of the learning period for campaigns).
Want this done for you? We book qualified sales appointments on a Pay-Per-Result basis — you only pay for calls that actually land in your calendar.
The next 24 hours: the 30-lead callback test
The whole triage, tonight, for nothing, with no agency involved.
- Pull your last 30 paid leads. Thirty leads, not thirty days. Record two timestamps each: when the enquiry landed, and when a human first tried to contact them.
- Ring all 30 yourself. Not your setter, not the agency. You. Two attempts each, at different times of day.
- Score each lead on three yes/no questions. A: did they answer? B: do they remember making the enquiry? C: do they have the problem you solve and the means to pay?
- Work out your median first-response time from step 1. Median, not average — one lead answered in 40 seconds and nine in three days averages to something that sounds acceptable and is not.
Budget three hours. The number that usually stops people cold is not A, B or C: it is the median response time.
What the three scores tell me to fix
Each pattern points at one stage. This is the table to act on.
| Pattern across your 30 | What it means | Where the money went | First move |
|---|---|---|---|
| A ≥ 60%, B ≥ 80%, C ≥ 50% | Traffic and reachability are both fine | The booking and closing conversation | Leave the ad account alone. Work on show rate and the call itself. |
| A < 30%, B high | Contactability, not targeting | Real buyers you never reached | Response time and attempt count. This is the top row of the leak table. |
| A high, B < 50% | Click quality or incentive mismatch | People who converted on the lead magnet, not the offer | The ad promise and the landing page, not the budget. |
| A high, B high, C < 25% | Targeting and qualification | The wrong audience, efficiently reached | This is the one case where pausing and rebuilding the audience is correct. |
| Median first response > 24 hours | Timing beats everything else you found | Decay between enquiry and first call | Fix this before you interpret any other row. |
If we can’t make you money, we don’t deserve yours.
Pay-Per-Result pricing — performance-based alignment.
The next 7 days: repair the one stage the test found
Fix one stage. Fix three at once and you will never know which worked. Google Ads or Meta, the order is the same: spending money on ads but not getting a return is almost always an ad spend not converting problem, not a traffic-volume one.
If it was contactability (A low): put a first touch on every new enquiry inside five minutes, seven days a week, and run at least six attempts across call, SMS and email over 14 days. The Lead Response Management study — InsideSales.com data on 15,000+ leads and 100,000+ call attempts across six companies and three years, analysed by Professor James Oldroyd — reports “a staggering 21-fold decrease in the odds of qualifying a prospect if the response time stretched from 5 to 30 minutes” (Lead Response Management study). Read it for what it is: an older, dialler-vendor-funded dataset measuring phone qualification by outbound teams, not close rate. The direction is well replicated; the magnitude is not a promise. What the five-minute number does and does not say is unpacked in the five-minute rule and the response-time research behind it, and the levers in how to increase contact rate.
If it was click quality (B low): match the landing page promise to the ad promise word for word and drop the incentive attracting people who forget they asked. If it was targeting (C low): rebuild the audience before the creative. If everything scored well: the leak is the conversation — measure show rate first.
The honest cost: diagnosis is three hours. Sustaining a five-minute first response is the expensive part — someone reachable at 7pm on a Sunday, a dialler that actually fires, and weekly transcript review. That is a part-time role plus tooling, and where most owners hire, automate, or quietly go back to replying next morning. AI appointment setting holds that line without the headcount; we have booked 50,769+ AI-booked sales appointments since 2017 off 1M+ leads generated, and we are paid on booked qualified appointments rather than a retainer — a different risk shape to your current arrangement, not automatically a better one.
When this is not a marketing problem at all
Three cases where no agency, including us, is the answer. Your unit economics: break-even return on ad spend is 1 divided by gross margin — 1.67x at a 60% margin, 2.5x at 40%. If your deal size, margin and a realistic close rate cannot carry your cost per acquisition at any plausible conversion rate, the offer is the problem and the conversation is with your accountant, not a media buyer. Billing, invalid traffic and account suspensions go to the platform, not a consultant. Disapproved ads in a restricted category — credit, health, employment, housing — are a policy and possibly a legal question for your own adviser. This page is general information, not legal or financial advice.
Should I pause my ad spend while I work this out?
No, with two exceptions. Pause if the 30-lead test returned high A, high B and low C: you are efficiently buying the wrong people. Pause if you cannot fund next month — a cash decision outranks everything here. Otherwise keep the account running while you fix the stage you found. Pausing when the fault is contact rate destroys the only asset you have — a supply of fresh enquiries — and buys a fresh learning period on the way back in. If you need to know whether the spend does anything at all, the clean test is a holdout, not a full stop: how to run an incrementality test on lead gen spend sets out the geo-split version.
The fix that stops it happening again
One line, seven numbers, every Monday: cost per click, landing page conversion rate, contact rate, booked rate, show rate, close rate, median first-response time. Ten minutes a week, and it makes the next bad month diagnosable in one glance instead of one argument.
In our own client work, fixing speed to lead alone has been worth roughly 3x on conversion and doubling contact rate roughly 2x. Those are our figures from our own campaigns, not research, and they do not multiply — 3x times 2x is not 6x, because faster response is part of how contact rate improves. Overlapping levers counted separately is how honest numbers turn into dishonest ones. Across clients who supplied before-and-after revenue we report a 7x average sales lift; our published methodology discloses the median is closer to 4x.
Frequently asked questions
Should I pause my ads while I work out what’s wrong?
Only if the leads are reachable and real but unqualified, or if you cannot fund next month. Otherwise keep spending and fix one downstream stage, because restarting costs you learning time: Google Ads states it “can take up to around 50 conversion events or 3 conversion cycles for the bid strategy to calibrate to the new objective” (Google Ads Help).
How long should I give ads before deciding they’re not working?
Long enough to contain at least 10 conversions, and never less than six weeks below that volume. At one deal a month, a zero-deal fortnight occurs about 61% of the time with nothing wrong at all. Judge campaigns on conversion counts, not calendar days.
How do I tell whether it’s the ads or my sales follow-up?
Ring your last 30 leads yourself and score answered, remembers enquiring, and genuinely qualified. High on all three means the traffic is fine and the leak is your follow-up or your call. Low answer with high recall means you bought real buyers and never reached them — a timing problem, not a targeting one. The Lead Response Management study, on 15,000+ leads and 100,000+ call attempts, reports “a staggering 21-fold decrease in the odds of qualifying a prospect if the response time stretched from 5 to 30 minutes” (leadresponsemanagement.org); it is an older, vendor-funded dataset measuring phone qualification.
Is a 0.4x return on ad spend normal or is something broken?
It is below break-even for almost any business. Break-even return on ad spend is 1 divided by gross margin: 1.67x at a 60% margin, 2.5x at 40%. Anything under that is losing money on every unit of spend, so the question is which stage, not whether.
My agency says the ads are fine. How do I check?
Ask for three numbers they cannot fudge: cost per click, landing page conversion rate, and the percentage of leads contacted within five minutes. The first two are theirs. The third is almost always yours, and it is usually where the answer is.
Can I fix this without hiring anyone?
The diagnosis, yes — three hours. Sustaining a five-minute response seven days a week is where do-it-yourself breaks. Below roughly 30 leads a month you can carry it personally. Above that, something has to answer when you cannot.
Pay-Per-Result appointments
See if we’re a fit
We book qualified sales appointments for you and you pay on results, not retainers. Our booking page asks a few quick questions so you find out in two minutes whether that model suits your business.
- 50,769+ appointments booked without cold calling.
- Pay-Per-Result pricing — you pay for booked, qualified calls.
- Pick your own time on our live calendar, no phone tag.