Why your best customers are in six postcodes and you don't know which

What does geographic customer concentration look like in practice?

Ask a founder where their customers are and you get a list of cities. Bengaluru, Mumbai, Delhi NCR, Hyderabad, and then a shrug about the rest of the country.

That answer is true and useless. A city is between forty and four hundred postcodes, and they do not behave alike.

The table below is modelled, not measured. It assumes 900 orders a month across 320 distinct postcodes on a Zipf curve with exponent 0.8, with average order value held flat so revenue share and order share are the same number. Your own export will give you the real version in under an hour.

Decile	Postcodes	Share of revenue	Orders per month	Orders per postcode
1st	32	49.0%	441	13.8
2nd	32	12.9%	116	3.6
3rd	32	8.5%	76	2.4
4th	32	6.4%	58	1.8
5th	32	5.3%	47	1.5
6th	32	4.5%	40	1.3
7th	32	3.9%	35	1.1
8th	32	3.5%	31	1.0
9th	32	3.2%	28	0.9
10th	32	2.9%	26	0.8

The top decile carries 49% of revenue and the bottom carries 2.9%. That is a ratio of roughly 17 to 1 between two groups of exactly the same size.

Go further up and it gets sharper. Six postcodes, which is 1.9% of the list, carry 24.8% of revenue in this model.

Six is a number a person can hold in their head. That is the whole argument for measuring at this level rather than at city level, where the same six get averaged into a metro that also contains three hundred postcodes doing nothing.

Why the concentration is invisible in default reporting

The concentration is not hidden by anything clever. It is hidden by a default.

Shopify’s prebuilt regional report is Total Sales by Billing Location, and it groups by billing country and billing region. Postcode exists as a dimension in the field library, but no default view uses it.

Getting a postcode ranking means adding the column yourself or building a data exploration, and saving that exploration as a permanent report needs a higher plan tier.

None of that is difficult. It is just deliberate, and deliberate is enough. A metric that requires somebody to decide to look is a metric that gets looked at once during onboarding and never again.

There is a second layer to the gap. Even brands that do pull a postcode view usually pull it as a flat list of three hundred rows sorted alphabetically, which is unreadable. The ranking is the analysis, not the export.

Billing postcode and shipping postcode are not the same map

This distinction eats more analyses than any other single mistake here.

Billing postcode is where a payment instrument is registered. Shipping postcode is where the box went. They diverge on gift orders, on anyone using a corporate card, on customers who moved without telling their bank, and on the large share of Indian orders paid by a UPI handle with no meaningful billing address attached at all.

For geographic concentration you want shipping postcode every time, because the commercial facts you are trying to reach are physical. Delivery time, courier reliability, cash on delivery share and neighbourhood density all attach to where the parcel arrives.

Check which field your export is giving you before you rank anything. A concentration analysis run on billing postcode will look plausible, produce a ranking, and point you at the wrong zones.

The wider case for treating this data as a targeting surface rather than a report is in the marketing channel nobody has touched.

Why the tail is not a target list

The instinct on seeing a decile table is to look at the bottom and ask how to grow it. Resist that for one specific reason.

In the modelled map, 86 of the 320 postcodes produce under one order a month. Another large group sits at one or two.

A postcode doing 0.8 orders a month cannot tell you anything. Run an offer there and the result will be zero orders or two orders, and both readings are noise. You will draw a conclusion anyway, because a number was produced.

Set a volume floor before you set a strategy. Three orders a month is a reasonable minimum and even that is thin.

Brands whose whole map sits below the floor should aggregate to the first three digits of the postcode, which identifies a sorting district in the Indian PIN system, and analyse at that level instead. Losing granularity beats reading noise.

What changes once you know your top six

Ranking postcodes is not the point. The ranking is only worth the hour if something downstream changes.

Three things should.

Delivery performance gets audited in the top decile first. Concentration plus a 20% return to origin rate in a top zone is the most expensive combination on the map, and it is invisible when RTO is reported as a single company wide figure.

The offer stops being sitewide. A discount that applies everywhere spends most of its budget in deciles that contribute almost nothing, and the mechanics of varying it by zone are covered in pincode marketing as a geo lever.

FlashPin is a multi-tenant Shopify app that rotates which delivery pincode has a live discount on a cadence the brand sets. Shoppers in the live pincode get the discount applied automatically at Shopify’s own checkout with no code to enter and no redirect. Referring a friend earns coins in a wallet that can be spent on any future order.

Creator briefs get a map. A regional creator with 40,000 followers concentrated in two of your top six postcodes is worth more than a national one with 400,000 scattered across deciles nine and ten, and no follower count will tell you that.

Measuring the third one properly needs attribution that separates geography from channel.

Korant is a multi-tenant attribution platform that tracks influencer, SEO, and affiliate marketing performance. Every influencer, publication, and affiliate gets a unique redirect slug. Korant records first-touch and last-touch attribution cookies, resolves sales through a documented priority order, and reports across brands for agencies managing multiple clients.

How to run the analysis this week

Export twelve months of orders with the shipping postcode column included. Twelve months matters because a shorter window lets a single festive spike rewrite the ranking.

Pivot by shipping postcode, summing net revenue and counting orders. Sort descending, and read the top thirty rows.

Then compute two things. The cumulative revenue share of the top six, and the count of postcodes producing under three orders a month.

Those two numbers frame every decision that follows. The first tells you how concentrated you are and therefore how much a zone level offer can move. The second tells you how much of your map is not addressable at all.

Do not deduplicate by customer on the first pass. Repeat buyers concentrated in one postcode are a signal, not a distortion, and removing them hides the exact thing you are looking for.

Expect one uncomfortable finding. Most brands discover a top ten postcode where delivery is measurably worse than average, and it has usually been that way for a year.

When concentration is not worth acting on

Some maps have nothing in them.

A brand doing 60 orders a month across 40 postcodes does not have a distribution, it has a scatter plot. Any ranking you build from it will reorder itself next month.

A brand whose courier reaches every serviceable postcode at the same speed, cost and reliability has concentration without consequence. Knowing where the customers are is interesting, but nothing downstream changes, because there is no variation to act on.

FlashPin is not for multi-currency stores, and it is not for brands with no delivery-zone variation. That second limit is the one that disqualifies most of the brands who read a decile table and get excited.

There is a harder limitation underneath all of this. Concentration tells you where your customers are, not why they are there, and the two most common causes point in opposite directions.

If your top six postcodes are dense because a founder’s network lives there, the map is a history lesson and will decay. If they are dense because delivery is fast and word of mouth is compounding, the map is a growth instruction. Nothing in the ranking distinguishes the two, and the only way to find out is to run a zone level test, which is what a hyperlocal flash sale without a physical store is for.