Finding out why people abandon before they pay

Twenty-two people asked about the kurta this month. Four bought. The other eighteen read your reply, maybe sent a thumbs up, and disappeared, and you've been guessing at why for three weeks.

Stop guessing and ask them. One message, one specific question, sent within about a week, to the people who went quiet and also to the people who bought. You'll have real answers by Friday.

And before you rebuild anything: for a shop that sells through DMs, the drop is almost never a cart. It's the delivery cost arriving after the price, or a payment method the buyer doesn't have, or a reply that came nine hours late. Those four things account for most of the eighteen, and each is fixable in an afternoon.

You probably don't have a cart problem

Cart abandonment as a concept comes from big stores with a checkout flow, a basket icon and a payment page. Most small sellers don't have that shape. Somebody messages, you reply, and somewhere in the next four messages it either happens or it doesn't.

So the useful question isn't "how many carts were abandoned". It's "at which message did people stop replying", and that's countable by hand.

While we're here: you'll see a lot of confident statistics about abandonment rates. Nearly all of them come from large stores in specific countries, and most of the ones circulating on blogs have no source you can check. They will not describe your shop and you shouldn't set targets against them. The same is true of most published conversion figures, and what conversion rate you should expect makes that case in detail.

The five places people actually stop

In rough order of how many people they cost a small seller.

1. The price, in silence. They asked, you answered, nothing. Sometimes the price is genuinely too high. More often the reply gave them nothing to do next, so there was no moment of decision.

2. The delivery charge. They'd accepted ₹899 and then found out about ₹120, or ₹250 outside the state. It isn't the amount. It's that the total moved after they'd agreed to it, which feels like being caught out.

3. The payment method. You take bank transfer, they wanted UPI. You take card, they wanted cash on delivery. In much of India, Nigeria, Indonesia and the Philippines, cash on delivery isn't a fallback, it's the default expectation, and a shop that doesn't offer it loses a chunk of buyers who never explain why.

4. The delivery time. Four to six days is fine if you say it up front and fatal if it appears at the end.

5. Something unanswered on the page. Will it fit. Can I return it. Are you real. Each of those sends people away silently, because nobody messages a stranger to say "I don't trust you."

Find out where before you find out why

One month, one tally. This is the whole instrumentation and it costs nothing.

Every time someone enquires, put them in one of these buckets when the conversation ends:

Bucket	What it means
Went quiet after the price	Price or a dead-end reply
Went quiet after delivery cost	Total moved
Asked about payment, then quiet	Rail mismatch
Asked about timing, then quiet	Too slow
Said no, gave a reason	Free information, write it down
Bought

Thirty conversations is enough to see the shape. If twelve of them cluster in one row, that's your problem and you don't need any further analysis.

Do it in a notes app if a spreadsheet feels like too much. The point is that you're counting instead of remembering, because memory over-weights whichever conversation annoyed you most.

Ask the people who bought, not just the ones who left

This is the part that changes everything and almost nobody does it.

People who didn't buy are hard to reach. They feel awkward, they don't owe you anything, and a decent share never reply. People who did buy are delighted to talk to you, and they hit exactly the same objections on the way in.

So ask them, in the message a day or two after delivery:

Quick one, and be honest, was there anything that nearly made you not buy? Doesn't offend me at all, it's genuinely useful.

The answers are specific and slightly embarrassing and enormously useful. "The delivery to Surat was more than I expected." "I wasn't sure if it was real cotton." "I nearly didn't because you took a day to reply and I thought the account was dead."

Every one of those is a thing eighteen other people hit and left over. You just got told about it by somebody who pushed through anyway.

The message for the ones who left

Send it about a week after they went quiet, once, and let them go afterwards.

Hi Priya, no worries at all if you've decided against it, honestly. If you've got ten seconds, was it the price, the delivery cost, or just timing? Helps me a lot to know. Either way, thanks for asking.

Three things make it work. It gives permission to have said no, which removes the reason they went silent. It offers three specific options, which is far easier to answer than an open question. And it asks for ten seconds rather than a conversation.

The reply rate to this is much higher than to a chasing message, and some of the replies turn into orders, which is a side effect rather than the goal. There's a full sequence for the messages before this one in how to follow up without being annoying.

What people say and what they mean

Two translations worth knowing.

"It's a bit expensive" usually means "I'm not sure it's worth it", not "I can't afford it". The fix is evidence rather than a discount: a customer photo, a measurement, the fabric weight, how long it lasts. If you drop the price and they still don't buy, it was never the price.

"I'll think about it" almost always means there's an unanswered question they're not asking. Ask what it is, directly and lightly: "Course. Anything you want to know about the fit or the fabric before you do?"

"Do you have it in stock?" asked after everything else is settled usually means "are you a real shop". Answer with a photo of the actual item on your actual table, today. It closes more sales than any wording.

The four fixes, cheapest first

Put the delivery cost in the first message. Free, takes one edit to your saved reply, and it's the single highest-return change on this list. You will lose a few people earlier in the conversation, which feels bad and is actually good, because you're losing them before you've spent twenty minutes on them. The people who continue have already accepted the real total.

If you don't know what your delivery should cost, that's a separate and very fixable problem, and how to work out delivery charges has the arithmetic.

Add the payment method your market actually uses. In India that means UPI, and a plain UPI ID needs no gateway, no setup fee and no website. In Nigeria it's a transfer. In Kenya it's an M-Pesa till. In much of Southeast Asia it's cash on delivery. You don't need a card processor to fix this, you need the thing your buyers already have on their phone.

Answer faster during two windows. You don't have to be available all day. Pick the two times most enquiries land, usually late morning and around 9pm, and clear the queue then. Consistency beats availability.

Put the answers on the page. Measurements, fabric, delivery cost, delivery time, what happens if it doesn't fit. Every question you answer on the page is a conversation that doesn't need to happen, and it catches the people who'd never have messaged at all.

Cash on delivery fixes one problem and creates another

Offering cash on delivery will visibly reduce abandonment in markets where it's expected, because it removes the trust question entirely. The buyer risks nothing.

It moves the risk to you. Some parcels come back refused, and you've paid for the trip twice. The refusal rate varies wildly by category, city and price point, and anyone quoting you a single number for it is guessing. Measure your own over 50 orders and you'll have a figure that's actually about your business.

Two things that reduce refusals without much effort: confirm by message the day before delivery, and take a small advance on higher-value orders so there's something at stake. A ₹100 advance on a ₹899 order changes the behaviour of the people who order on a whim.

Sneha's kurtas, in Ahmedabad

Sneha sells block-printed cotton kurtas from Ahmedabad. ₹899 for the standard, ₹1,250 for the longer one, ₹120 delivery within Gujarat and ₹250 elsewhere in India. Instagram and WhatsApp, UPI or cash on delivery.

She counted one month. Thirty-one enquiries:

Where they stopped	Count
Went quiet after price	9
Went quiet after delivery cost	11
Asked about payment, quiet	2
Asked about sizing, quiet	3
Bought	6

Eleven out of thirty-one dropped at the delivery cost. It was the biggest single bucket and she'd have guessed price.

She changed one thing: the delivery cost went into her first reply, alongside the item price, with the outside-Gujarat figure named too.

The next month, 28 enquiries. Fewer people continued past the first message, which looked like a step backwards for about ten days. But 10 of the 28 bought, against 6 of 31.

Then she asked her buyers the "what nearly stopped you" question. Four of the ten mentioned the ₹250 out-of-state charge, and two said they'd assumed the kurta was polyester because the photos were bright. She added the fabric weight in grams to every listing and reshot two products in daylight.

The detail she found most annoying: three of the nine who went quiet after the price, when she finally asked, said they'd been waiting for her to reply to a follow-up question that she'd never received. Instagram had put their messages in the requests folder because they didn't follow her. She'd been reading that silence as rejection for months. Now she checks the requests folder every morning, which takes fifteen seconds. There's a lot more of this sort of thing in selling online in India.

What the numbers can't tell you

This is worth being straight about, because it's a real limit of any shop tool including Sailo.

Analytics can count people who opened your shop link and didn't order. It cannot see the WhatsApp conversation where the actual decision happened, and for most small sellers the decision happens in the conversation. So the dashboard will show you a drop and never the reason, and the reason is only ever available from a person.

Two more specifics. On manual payment rails, an order placed and not paid for looks identical to an abandoned one until you check your bank, and nothing automates that check. Sailo can't tell you a bank transfer or UPI payment arrived, only your bank can, which means "abandoned" and "paid, and I haven't looked" are the same row in your list until you do the looking.

And on the free plan, analytics only go back 7 days. That's enough for a one-month tally like Sneha's and not enough to see whether a change you made in March is still working in September. Pro holds a year, Business three. The tally in your notes app has no expiry date, which is the argument for keeping it regardless of which plan you're on.

Send five messages this week

Take the last five people who bought from you and send them the "was there anything that nearly stopped you" question. That's five minutes of typing and it's the highest-quality information about your shop that exists anywhere.

Then take the last five who went quiet, send the three-option version, and expect two replies.

Whatever comes up twice in those ten conversations is the thing to fix first. If your enquiry volume is too low for any of this to produce a signal, that's a different problem and it comes first, which is what getting your first ten orders is for.