Selling on Instagram in Pakistan
The inbox is the checkout, cash on delivery is the payment, and the ad's only job is to start a conversation. The operator's version, with The Gallery's result.
Most advice about selling on Instagram was written for a market with a card in every wallet and a checkout on every site. Pakistan is not that market. Here the buyer sees the post, sends a message, asks the price, confirms a size, and pays the courier at the door. Treat that as a defect and you will spend a year fixing something that was never broken. Treat it as the funnel and it sells.
Why does the sale close in the inbox?
Because trust does. A buyer here wants to know that there is a person behind the page, that the product exists, that the price is the price, and that the thing will arrive. A message answers all four in a minute. A checkout page answers none of them, and asks for a card that many buyers do not have or do not want to use online.
The inbox also does the selling a page cannot. Personalised products need a conversation: a name, a spelling, a date, a colour. A buyer who was undecided on the post decides in the reply. Instagram knows this, which is why its advertising can be optimised toward messages rather than clicks, and why a campaign built for a Pakistani brand should be.
What does cash on delivery do to your numbers?
It moves the risk to you. The order is not a sale until the courier has been paid, and a share of orders will be refused at the door, or never answered when the rider calls. That share is the number that decides whether the business works, and it is set by what happens in the inbox before dispatch, not by the ad.
Confirm every order before it ships: a message that repeats the product, the price, the address and the delivery window, and asks for a yes. Call the ones who do not reply. Ship nothing unconfirmed. The brands that lose money on cash on delivery are the ones that treat the first message as the order. The ones that make money treat the confirmation as the order and the first message as a lead.
What does the reply script look like?
Speed first. A buyer who messaged you also messaged two other pages, and the first clear reply usually wins. Then the script, which is not a sales pitch but a short set of answers in the order the buyer needs them.
| The buyer says | You reply with | Why |
|---|---|---|
| “Price?” | The price, the sizes or options, and the delivery charge, in one message | Three questions answered before they are asked; nothing to wait for |
| “Is this available?” | Yes, the options in stock, and the delivery time to their city | Confirms the product exists and that you know your stock |
| A screenshot of a post | The product name, the price and the one question you need answered to proceed | Moves from browsing to ordering in a single step |
| “Can I get it by Friday?” | A true answer, and the alternative if the answer is no | A missed promise is a refused parcel |
| Nothing, after the price | One follow-up the next day, with nothing new to say except that you are here | Many inbox sales close on the second message, not the first |
| “Yes, order it” | The full order read back, with the address and the amount the courier will collect, and a request for a yes | This is the sale. Everything before it was a lead |
Write the script down and give it to whoever holds the phone. The inbox is a sales desk, and a sales desk that improvises loses orders to one that does not. Saved replies on the platform make the first message near-instant; a person takes over at the first question the saved reply cannot answer.
When should you add a store?
Later than you think, and for a different reason than you think. A store does not replace the inbox here. Most buyers will still message. What a store does is hold the catalogue, so the inbox stops being asked what else you have, and it gives the brand a place that looks like a business to a buyer who has not yet decided whether to trust the page.
Add one when the inbox is answering the same catalogue question all day, when repeat buyers ask for a link, or when you start selling outside Pakistan to people who do pay by card. Keep it simple, keep the prices the same as the ones quoted in the inbox, and keep cash on delivery as the first payment option on it, because that is still how most of the orders will arrive. The piece on store choice is at /blog/shopify-store-or-custom-site, and the Shopify page is at /services/shopify-stores.
What kind of ad starts a conversation?
Not the kind that says shop now. An ad that ends in a checkout is asking for a decision the buyer is not ready to make. An ad that ends in a message is asking for a question, and questions are cheap to ask. The creative has to earn the question.
- Show the product being made, personalised or packed. Process is proof that it exists and that a person is behind it.
- Put the price in the creative, or the starting price. A buyer who messages knowing the price is halfway to confirming; one who messages to ask it often leaves when told.
- Ask the one question the buyer would ask you: what name, which colour, which date. The reply is the message.
- Use the platform's message objective, so the campaign is optimised toward conversations rather than clicks, and report conversations, not link taps.
- Run a tight set of creatives, cut the losers weekly and rebuild the winners. More creative does not help if nobody is reading the inbox.
What happened for The Gallery?
The Gallery sells personalised newborn wear in Pakistan. Every product needs a name, a spelling and often a date, so the checkout was always going to be a conversation, and a campaign that pushed traffic at a product page would have measured the wrong thing.
We built ten creatives, each made to start a conversation rather than close a cart, and optimised the campaigns toward messages. The sale was tracked where it happened, in the inbox, and reported weekly against what had been agreed. Every campaign to July 2026 returned $5.20–$6 back for every $1 spent on ads. Not an average across one good quarter; every campaign. The spend behind the figure is not published, which we would rather say plainly than round away. The case is at /work/gallery.
What should you measure?
Orders delivered and paid, and what each one cost in advertising. Everything else is a step toward that or a distraction from it.
| Measure this | Not this | Because |
|---|---|---|
| Conversations started per campaign | Likes, saves, reach | A like has never paid a courier |
| Confirmed orders, and the share of conversations that became one | Link clicks | This is where the inbox either works or does not |
| Parcels delivered and paid, and the share refused at the door | Orders placed | Cash on delivery means an order is a promise until the rider is paid |
| Advertising cost per delivered order | Cost per thousand views | The only cost that can be set against margin |
| Reply time, in minutes | Follower count | Speed decides who wins the buyer who messaged three pages |
Keep a sheet. One row per order: the campaign it came from, the day it was confirmed, the day it was delivered, whether it was paid. That sheet is the report; the platform's dashboard is an input to it. An agency that reports from the dashboard alone is reporting on advertising. You are running a shop.
If we run it, the ad account stays yours, the spend stays on your card, and the fee is flat rather than a share of the spend, so telling you to pause costs us nothing. The longer piece on why the funnel is shaped this way is at /blog/the-dm-first-funnel-selling-on-instagram-in-pakistan, and the Pakistan page is at /pakistan/meta-ads-agency-pakistan.