Sequencing a quick commerce city launch without burning cash
Going live in ten cities at once feels like ambition. It is usually the most expensive way to learn what one city would have taught you in six weeks.
- Prove the unit economics in one city before buying presence in ten. The learning is transferable, the wasted stock is not.
- Pick the first city on demand density and your own supply reliability, not on market size.
- Expand on evidence: repeat rate, sales per facing and availability held above target for a full cycle.
- Thin inventory spread across many cities produces poor velocity everywhere and weakens your next assortment review.
The pitch is familiar. The platform offers to take you live across a wide footprint, the category manager talks about visibility, and the internal conversation shifts to how many cities the brand will be in by the end of the quarter.
City count is a satisfying number. It is also close to meaningless as a measure of whether the channel works, and chasing it early is how brands end up with stock spread thin, poor velocity everywhere and a difficult assortment review six months later.
What you are actually trying to learn
A first launch is a test, and the test has a specific question: does a unit of this product, sold through this channel, in a realistic operating condition, make money after everything.
Everything means platform margin, listing and promotion costs, the ads you needed to get discovered, the cost of holding stock close enough to replenish, damages, expiry on anything perishable, and the returns or write offs the channel generates.
That question is answerable in one city. Answering it in ten costs ten times as much and produces a muddier answer, because the variables move independently and you cannot tell which city taught you what.
Choosing the first city
The instinct is to pick the biggest market. The better criterion is where you can hold availability, because availability is the precondition for every other number being meaningful.
Ask three questions. How far is your stock from those dark stores, and how fast can you replenish when a SKU runs down. Do you already have demand signal in that city from your own site or marketplace sales, which tells you the product has a buyer there. And is the category competition in that city so intense that your early data will reflect the promotional environment rather than your product.
A mid sized city near your warehouse where you already see organic demand is usually a far better first test than a metro where you will spend the entire launch period out of stock and fighting for visibility.
The availability trap
Here is the sequence that ruins most launches. Stock goes in. The platform promotes the new listing. Demand spikes. The spike drains inventory faster than the forecast assumed. Replenishment takes two weeks because nobody planned for the spike. The listing sits out of stock through the period when the platform was measuring it.
When stock returns, the velocity data on record is poor, the algorithm has learned not to surface the product, and the brand concludes quick commerce does not work for its category. What actually happened is that a supply chain designed for steady demand met a demand curve that was not steady.
Plan the first replenishment before you go live, not after. Assume the launch spike is real and temporary, and hold buffer stock close enough to respond within days.
What has to be true before city two
Set the gate in advance, in writing, so that expansion is a decision rather than a mood.
Availability held above your target for the majority of the period, because without that nothing else is readable. Sales per facing at or above the category median, which you should ask your category contact for rather than guess. Evidence of repeat purchase, since quick commerce is a replenishment channel and a product nobody buys twice will not sustain shelf space. And a contribution margin that works at the promotional intensity you actually ran, not at list price.
If three of the four hold, expand. If only one or two do, fix the gap in the city you understand before adding cities you do not.
Expand along supply lines, not the map
When you do expand, the next city should be chosen by how easily your existing supply chain reaches it, not by market size ranking.
Two cities served by the same warehouse are far more than twice as easy as two cities served by different ones. You reuse the same replenishment rhythm, the same buffer, the same team routine and the same escalation path. Density in a region compounds; scattered presence does not.
This is also how you avoid the most common late stage failure, which is a brand live in fifteen cities with a supply chain built for four, discovering that availability is now unmanageable in all of them.
Say no to the wrong kind of yes
The hardest part is declining a national listing offer, because it feels like turning down growth and the platform frames it that way.
It helps to remember the asymmetry. Breadth of listing costs the platform almost nothing and helps it fill assortment. The inventory, the working capital and the velocity scorecard are entirely yours. A brand that goes wide and performs poorly does not get a neutral outcome, it gets a data trail that argues against it in the next review.
A phased plan with agreed evidence at each step is a perfectly professional answer, and category managers deal with it routinely. The brands that struggle are the ones that said yes to everything and then could not supply any of it.