01 · The problem
Why scaling breaks cash on delivery specifically
In a prepaid business, doubling ad spend roughly doubles revenue and roughly doubles cost. The relationship is close to linear and the money arrives immediately.
COD is not linear, for two reasons that compound.
Your delivery rate moves when your audience widens. The first thousand buyers are the people your ad found most easily: the ones most interested. Widening the audience reaches people who are less sure, and less sure means more refusals at the door. You pay for that twice, outbound and return.
Your money arrives later than you spend it. Ads are paid today. COD is collected after delivery, then reconciled, then paid out. Every euro of extra ad spend widens that gap before it narrows it.
A COD business can be profitable on every single order and still go broke. The order is profitable; the timing is what kills you.
02 · The prerequisite
Find your ceiling before you spend a euro more
You cannot scale a number you have not measured. Before increasing budget, you need at least 100 delivered orders on one product in one country. Not 100 leads, not 100 confirmed. Delivered.
From those you need three figures, which you should be able to state from memory:
- Confirmation rate. Leads that became confirmed orders.
- Delivery rate. Confirmed orders that were actually accepted and paid for.
- Cost per delivered order. Total ad spend divided by delivered orders. Not cost per lead. This is the only ad number that means anything in COD.
If you do not have these, you are not ready to scale, you are ready to gather data. Sellers who skip this step interpret every later problem as an ad problem, because the ad dashboard is the only place they are looking.
03 · Ad spend
How to actually raise the budget
Double, then wait for delivered orders
Not 20% increments, which are too slow to learn from, and not 5x, which destroys the data. Double the budget and then wait until you have another 100 delivered orders at the new level before judging anything.
Compare delivery rate, not revenue
Revenue always goes up when you spend more. That tells you nothing. The question is whether the same percentage of parcels is still being accepted. If it held, the ceiling is higher than where you are. If it dropped, you found it.
If the delivery rate fell, go back and fix, do not push
A falling delivery rate at higher spend means the wider audience is worse, or the ad is now promising something the product does not deliver. Both are fixable. Neither is fixed by spending more.
Re-establish the new baseline before doubling again
Each doubling needs its own 100 delivered orders. It feels slow. It is far faster than scaling into a wall and spending three months unwinding it.
04 · Money
The cash gap, and how to size it
Between spending on an ad and receiving the COD from the order it generated, several things have to happen: the lead is confirmed, the parcel is packed, the courier delivers, the buyer pays, the courier remits, and the payout runs. On our network payouts go out weekly, every Friday, to your own bank account.
So at any moment you are carrying roughly one to two weeks of ad spend plus the stock currently in transit or on the shelf. Every time you double the ad budget, you double that carry.
Practical rule: before doubling, make sure you can fund two weeks of the new budget from cash you already have, without counting on money that has not landed. If you cannot, do not double. Increase by half instead and take longer.
The stock half of the problem
Scaling ads without scaling stock produces the other failure mode: you sell what you do not have, lead times run out, and you either cancel confirmed orders or take a delivery-rate hit from delays. Order the next batch when you are at roughly three weeks of cover, not when you run out. Storage on our warehouse network is free, so holding a buffer costs you nothing but the cash tied up in it.
05 · Geography
Adding countries, in the right order
Once a product works in one market, geography is usually a better scaling lever than budget, because you are reaching a fresh audience rather than a worse slice of the same one.
The sequence that works:
- Start where you already have stock. From the Romania hub you can reach nine countries without moving a pallet. Adding Bulgaria or Greece to a working Romanian campaign costs you a translation and an ad set, not an import.
- Add the free-return routes while you are learning. Hungary, Czechia, Slovakia and Slovenia return free out of Romania, so a country that turns out not to work costs you the outbound leg only.
- Translate properly. The confirmation call already happens in the buyer's language. Your landing page should too. A machine-translated page converts worse and confirms worse.
- Treat each country as a separate test. Same 100-delivered-orders rule. A product that works in Romania is a hypothesis in Poland, not a certainty.
The per-country rates and transit times are on the courier comparison.
06 · Operations
What actually breaks when volume rises
Confirmation capacity
Twice the leads means twice the calls. If confirmation slows, parcels ship later, buyers cool off and the delivery rate drops for reasons that have nothing to do with your ad. Tell your partner before you scale, not after.
The return shelf
Returned stock needs checking, restocking and reselling. At low volume you can ignore it. At high volume ignoring it means a third of your inventory quietly sitting in a corner while you reorder more.
Stock cover
Supplier lead times do not shrink because you got busy. Reorder on a schedule tied to your run rate, not on the day you notice the shelf looks empty.
Your own attention
Three products across four countries is twelve campaigns, twelve delivery rates and twelve stock positions. Most solo sellers hit their personal ceiling well before their market ceiling.
07 · Limits
The signals that mean stop
Scale down or pause when you see any of these, and do not argue with them:
- Delivery rate falls two doublings in a row. You are past the audience that wants this product.
- Cost per delivered order rises while cost per lead falls. You are buying cheaper, worse traffic. This is the most commonly missed signal because the ad dashboard looks like it is improving.
- You are funding ads from money that has not arrived yet. Stop immediately. This is the failure that ends businesses rather than campaigns.
- Returns are outgrowing your ability to process them. Fix the shelf before adding more volume to it.
None of these mean the product is finished. They mean this configuration is at its limit, and the next move is a new country, a new creative or a new product, not a bigger budget on the same one.
FAQ
Questions people ask
After roughly 100 delivered orders on one product in one country, with a stable delivery rate and a known cost per delivered order. Before that you do not have enough data to tell a real result from noise, and any problem that appears later will get blamed on the wrong thing.
Double, then wait for another 100 delivered orders before judging. Smaller steps are too slow to learn from and larger ones destroy the data. Judge each step on whether the delivery rate held, not on whether revenue went up, because revenue always goes up when you spend more.
Because widening the audience reaches people who are less interested. The first buyers are the easiest ones to find. Beyond them you reach people who are less sure, and less sure means more refusals at the door. It usually indicates the ceiling of that audience rather than a broken operation.
Enough to fund about two weeks of your new ad budget plus the stock in transit and on the shelf, from cash you already hold rather than money still to arrive. Every doubling of ad spend doubles that carry, which is the failure that ends businesses rather than campaigns.
Usually yes, once a product works. Raising budget reaches a worse slice of the same audience, while a new country reaches a fresh one. From a Romania warehouse you can reach nine countries without moving stock, so the marginal cost of testing a second market is a translation and an ad set.