Where am I losing sales? A count that shows which stage is leaking

To find which stage of your sales process is leaking, take every enquiry from a quarter and mark the furthest stage each one reached. Count how many arrived at each stage and compare it with the one before. The biggest fall is not always the problem, because some falls are normal. The leak worth fixing is the one that loses buyers who looked ready.

Why not just look at the pipeline you have now?

The deals open today are the survivors. The ones that left are gone from the screen, and they hold the answer. Our post on tracking a pipeline and finding where deals get stuck looks at what is open. This one looks at everything that came in over a quarter, open or not, and asks where each one stopped.

How do I run the count?

List every enquiry from the last three months, one per row. Add one column: the furthest stage reached. Use plain stage names and add a final row for enquiries you cannot trace. Then count.

For example, imagine an invented laptop repair and annual-maintenance firm in Palakkad that serves small offices and shops. The owner is sure the quotes are too dear. His count for a quarter looks like this (illustrative figures):

Do not take these as targets. They show the method: each row says how many got there, and the last column is the gap to the next row.

The owner's instinct says the quote stage, where 21 of 40 walked away. The count says something else.

How do I read which drop is the leak?

Use three tests, in this order.

Were they ready when they left? The 12 who would not book a visit may be price-checkers, and that is a normal loss. The 8 who did get a visit and then never received a quote came to you, and you did not do the next step. The 30 who got no call back in the day never even met you. None of the three costs a rupee of advertising to repair.

Whose side is the loss on? Where nothing was sent, the loss is yours: a call not returned, a quote not written. Where something was sent and the buyer said no, it is the buyer's decision, and the repair is price, proof or the way you follow up.

What is it worth? Multiply the people lost by your usual close rate from that stage, then by an average job. Suppose, illustratively, that one in three of the people who spoke to the firm go on to approve a quote, and an average job is ₹6,000. Thirty callers who got no call back are then worth about ₹60,000 a quarter. The 8 quotes never written come to about ₹16,000. Pick the repair with the bigger rupee value, not the bigger count and not the bigger percentage.

In this made-up Palakkad count the biggest percentage drop was at the quote, but the stage worth fixing first was the very first one. Enquiries came in the evening after the shop shut, and nobody rang until the next afternoon. The quote was fine. The repair was a rota for the evening phone, which cost one technician about two hours a week of overtime, about ₹3,000 a quarter. Next quarter the firm reaches 70 spoken-to out of 95 enquiries, better, though the quote stage still loses half of those who receive one.

What are the usual causes of a leak?

Each stage tends to leak for its own reason, so match the cause to the stage before you act.

• Enquiry to reply. Slow answers. A number nobody watches on Sunday. A missed call that is never returned.
• Reply to visit. No clear reason to book now, or the buyer is still deciding what they want.
• Visit to quote. The person who visited cannot price on the spot, so the work waits for an hour the owner does not have.
• Quote to order. Price. Trust. A partner who has not yet agreed.

If you cannot tell which applies, ring five of the people who dropped and ask. Say you are trying to improve, not chasing. Some will tell you, and one honest answer is worth the five calls.

What if my records are a mess?

The first count will be rough, and that is fine. Mark the furthest stage by checking chats, call logs and the quote file, going back as far as the messages allow. Anything you cannot trace goes in a row marked "unknown". A high unknown count is itself a finding: you are losing enquiries before they are recorded.

Skip the exercise if you had fewer than thirty enquiries in the quarter. The percentages will swing on one or two jobs, so read each lost case instead.

When should I count again?

After you change something, and not before. If Palakkad starts returning every missed call the same evening, count the next quarter and see whether more people reach the technician visit. If nothing moved, the cause was somewhere else. Counting every few weeks only gives you numbers too small to read.

Put the second count beside the first and keep both. The change between them is worth more than either on its own.

Getting a second view

Start with the last ninety days of WhatsApp chats and write the furthest stage beside each name. It takes an afternoon. The sales process hub explains each stage. If you want someone else to read the count with you, see lead leakage analysis and tell us what you need.