Enquiries are down. Or enquiries are fine but sales are not. Or sales dipped last month and nobody can say why. Most owners know something is leaking, but not where. So they change the thing they can see, usually the adverts, when the real problem might be a page that broke on iPhones, a rep who stopped calling back on Fridays, or reminders that quietly stopped sending.
The fix is boring and reliable: put a number on every step from click to sale, and look for the step that changed.
Map every step
Write out your funnel as a list of stages. For a typical installer running ads to a landing page:
- People who saw the advert.
- Clicks.
- Landing page visits.
- Started the questions.
- Finished and became an enquiry.
- Contacted (a real conversation, not just a message sent).
- Qualified.
- Survey or appointment booked.
- Appointment went ahead.
- Quote sent.
- Sale.
- Installed and paid, without cancellation.
Your list may be shorter or longer. What matters is that every lead moves through the same named stages, every time.
Get a number for each step
The numbers live in different places, which is why most firms never see the whole picture:
- Steps 1 and 2: your ad platforms (Meta Ads Manager, Google Ads).
- Steps 3 to 5: your website analytics. In Google Analytics 4, a funnel exploration report shows how many people move from one step to the next and where they drop out.
- Steps 6 to 12: your CRM, but only if your team moves every lead through the stages properly. If half the leads sit in "new" forever, the bottom of your funnel is invisible.
The single most useful thing you can do is make sure each lead in the CRM carries its source (which campaign and advert) and is moved through every stage. That turns a pile of contacts into a funnel you can read.
Work out the rate for each step
Divide each step by the one before it. An example with made-up numbers for one month:
| Step | Count | Rate from previous step |
|---|---|---|
| Landing page visits | 4,000 | |
| Started questions | 1,200 | 30% |
| Finished (enquiries) | 480 | 40% |
| Contacted | 360 | 75% |
| Survey booked | 150 | 42% |
| Survey went ahead | 105 | 70% |
| Quote sent | 90 | 86% |
| Sale | 27 | 30% |
Now compare each rate with your own previous months. Industry benchmarks are of limited use here, because trades, offers and qualifying differ so much. Your own history tells you what normal looks like for you.
Diagnose the leak
Once you know which step dropped, the list of likely causes gets short.
| Where it leaks | Likely causes | What to check |
|---|---|---|
| Ads seen but few clicks | Tired creative, wrong audience | Ad frequency, new hooks, fresh video |
| Clicks but visitors leave | Page does not match the ad, slow on mobile | Open the ad and page side by side on a phone (landing pages) |
| Start questions but do not finish | Too many questions, a hard one too early | Drop-off by question (qualifying questions) |
| Enquiries but not contacted | Slow reply, wrong numbers, nobody chasing | Time to first reply (speed to lead) |
| Contacted but not booking | Objections not answered, no slots offered | Read the conversations and call notes |
| Booked but no-shows | Weak confirmation and reminders, slots too far out | Reminder sends, booking lead time (no-shows) |
| Surveyed but no quote | Poor qualifying, unsuitable homes | Surveyor feedback by lead source |
| Quoted but not sold | Follow-up stops, price shock | Follow-up activity after quotes (quotes that go quiet) |
| Sold but cancelled | Finance declines, expectations not set | Cancellation reasons |
Split the numbers before you trust them
A leak often hides in one segment while the overall rate looks fine. Split each step by:
- Lead source and advert. One campaign may bring plenty of enquiries that never book.
- Device. A page that works on desktop but breaks on one phone browser can cost you for weeks.
- Day and time. Weekend leads contacted on Monday behave very differently from leads contacted in minutes.
- Rep or surveyor. Different people convert at different rates at the same step.
- Area. Postcodes far from your base may book but not show, or show but not buy.
Look at what actually happened
Numbers tell you where. To find out why, look at real examples from that step:
- Fill in your own form on your own phone.
- Read twenty enquiry conversations, chat transcripts or call notes from the leaky step.
- Ask your surveyors about the last ten visits that went nowhere.
- Ring a few people who dropped out and ask, politely, what happened.
Twenty minutes of this often explains a drop that a week of staring at dashboards did not.
Fix the most expensive leak first
Work out what each improvement is worth in sales. Using the example above: raising the share of booked surveys that go ahead from 70% to 80% adds 15 surveys and, at the same rates below, around 4 extra sales a month, without spending a penny more on ads. A similar improvement at the top of the funnel has to survive every step below it before it turns into a sale.
Then change one thing, wait long enough for the numbers to settle, and check again.
A weekly routine
Put fifteen minutes in the diary every week:
- Check each step's rate against the last four weeks.
- Note anything that moved sharply.
- Split that step by source, device, day and rep.
- Read a handful of real examples.
- Agree one fix and who owns it.
Which numbers to put on your weekly dashboard is covered in our piece on the numbers to check every week. Tracing each sale back to the campaign that produced it is covered in our piece on tracking leads to sales.
Where we fit
We put tracking on every step of the funnels we build, and Installer Flow, our own CRM, carries each lead from the advert it came from to the sale. The reporting gives you dashboards on every number: pipeline, reps, cost per lead and cost per sale. If you know something is leaking but not where, book a video call and bring whatever numbers you have.
Sources
Questions owners ask
What is a good conversion rate for an installer landing page?
It varies too much by trade, offer, traffic source and how many questions you ask for a single figure to be useful. Your own history is the better benchmark. Track each step's rate week by week, and treat a sudden change or a slow slide as the thing to investigate, rather than chasing someone else's number.
Which leak should I fix first?
The one costing the most money, which is usually lower down the funnel than people expect. Raising the share of booked surveys that go ahead, or quotes that turn into sales, often adds more jobs than a better click-through rate, because every lead above that step has already been paid for.


