COD sales funnel: lead to delivery

Why does a single conversion number hide the real problem?

Because "3% of leads became paid customers" is the product of four separate conversion steps multiplied together, and a low number at the end could come from a weak first step, a weak last step, or a weak step somewhere in the middle that nobody is looking at closely. A store that reads one aggregate figure and concludes "we need more marketing" might actually have plenty of interested leads already coming in every day, and a confirmation process that loses most of them before they ever become a real order worth shipping. The aggregate number tells you clearly that something is wrong somewhere in the chain. It does not tell you what that something is, or where in the chain to start looking first before spending time or money on a fix.

What are the five stages, specifically?

Lead: someone asks about a product, whether on WhatsApp, Instagram, or a store page, before ever placing an order with any real commitment behind it. Order placed: they commit to buying, through checkout or by saying so directly in a message sent to the page. Order confirmed: they have given written agreement to the specific product, price and address as restated to them in a follow-up message. Order shipped: a parcel has been created and physically handed to a carrier for delivery to the customer. Delivered and paid: the carrier has handed it over to the customer and collected the cash on your behalf, closing the loop. Each stage has a clear entry and exit condition attached to it, which is exactly what makes it countable rather than a vague impression carried around in someone's head from memory.

How do you actually track this without new tooling?

At minimum, timestamp when each stage happens for every order that comes through: when the lead first messaged, when they placed the order, when confirmation was received in writing, when the parcel shipped out, when it was delivered and paid for in full. A spreadsheet with five date columns per order is genuinely enough to start tracking this meaningfully, no software purchase required to begin. What matters most is consistency: every order needs all five timestamps recorded the same way every time, or the funnel numbers become unreliable the moment some orders skip a stage in the data even though they did not actually skip that stage in reality on the ground.

What does reading the funnel actually look like?

In this example the two weakest transitions are lead-to-order at 55%, and delivered-to-paid at 73%, meaning 27% of shipped orders returned instead of converting into cash. Two different problems, two different fixes: the first is a sales and response-time issue sitting at the top of the funnel, the second is a delivery and confirmation-quality issue sitting much further down. An aggregate 24% conversion figure tells you neither of these things on its own, and points you at nothing actionable to actually change.

What should you actually do with this once you are tracking it?

Look at it weekly, and focus on whichever transition has the lowest conversion rate relative to what similar stores or your own past performance shows is realistically achievable for that specific stage of the process. Fixing the weakest link moves the overall number more than evenly polishing every stage a little at a time across the board. Once that link strengthens and stabilizes over a few weeks, a different stage will usually become the new weakest one in turn, and the funnel view tells you exactly where to look next instead of leaving you guessing at which part of the business needs attention this particular month.