Choose one way of reaching new customers, such as pamphlets, a society notice board, local ads on Google or Instagram, or a tie-up with a nearby business, and run only that for four to six weeks with a set budget. Give it a way to be counted (a code, a separate number, a question at billing), then compare the new customers it brought with what it cost. Keep what pays for itself; drop what doesn't.
Why it matters
Most small businesses try several things at once (a pamphlet drop, a boosted post, a discount on a delivery app) and then can't tell which one worked. The money goes, some new faces come, and the next decision is a guess again.
Testing one at a time is possible now because you have a baseline. You know roughly how many new customers come in an ordinary week, what an average bill is, and how many come back. Against that, a channel either moves the numbers or it doesn't. Before opening you had none of this, which is why the opening offer is planned differently.
Do this after your free channels are in order: a current Google profile, answered reviews and a list of regulars you can message. Paying to bring people to a profile with wrong hours wastes the spend.
How to do it
Write down your baseline
From your billing records or a simple count, note for the last four weeks: new customers a week (ask "first time here?" for a fortnight if you don't know), average bill, and how many first-timers came back within a month. That's what the test is measured against.
Pick one channel that fits your customer
Go where your customer already is. Options many local businesses try: pamphlets in newspapers or at society gates, a stall or notice at a residents' association event, a school or office tie-up, local-area ads on Google or a social platform, a listing on a delivery or booking app, or a referral offer for existing customers. Choose one.
Set the budget, period and target
Decide the spend (for example ₹5,000–₹15,000 as a ballpark for a small local test) and the period (four to six weeks, avoiding festival weeks that distort sales). Work out how many new customers it needs to pay back, using your margin per bill. If you haven't worked out margin, see working out break-even.
Make it countable
Give the channel its own trace: a code on the pamphlet ("Show this for a free chai"), a separate QR that opens your WhatsApp Business chat with a pre-filled message, a landing link, or simply a question at billing ("How did you hear about us?") with a counting sheet by the counter. Tell staff to ask every new customer, every time.
Run it and don't change anything else
During the test, hold your prices, hours and other promotions steady. Record new customers from the channel weekly. Note anything outside your control: rain, an exam week, roadworks.
Judge it on repeat customers, not just first visits
At the end, divide what you spent by the number of new customers it brought. Then check a month later how many came back. A channel that brings fewer people who return often beats one that brings crowds chasing a discount once.
| Illustrative test | Pamphlets at 3 society gates |
|---|---|
| Spend over 5 weeks | ₹8,000 |
| New customers counted with the code | 40 |
| Cost per new customer | ₹200 |
| Margin per average bill | ₹90 |
| Came back within a month | 14 |
| Decision | Keep, at the two gates that brought most codes |
In that example, one visit doesn't pay back ₹200, but customers who return a few times do. Whether that's worth it depends on how often your regulars come; that's why the return count matters as much as the first count.
Your result
One completed test with a clear keep, change or stop decision, and the next channel to try.
Baseline: new customers/week ___ avg bill ₹___ margin/bill ₹___ Channel tested: Dates: ____ to ____ budget ₹___ How it was counted: (code / QR / question at billing) New customers from it: ___ Cost per new customer: ₹___ Came back within a month: ___ Decision: keep / change / stop, because: Next channel to test:
Fill it in on your checklist, next to this topic. It stays in this browser.
Time and cost
- Time
- Four to six weeks per test, plus a month afterwards to count return visits. An hour to plan it, five minutes a week to record.
- Cost
- Your chosen test budget, often a few thousand to ₹15,000 for a small local channel (a ballpark; printing, ad and app rates vary by city). Counting costs nothing but staff discipline.
Ballparks to verify: prices and fees change by city and over time, so check current rates before you budget.
Common mistakes
Overlapping promotions
If pamphlets, an ad and a discount all start the same week, you can't tell which one worked.
Not counting at the counter
Without a code or a question at billing, you're left with a feeling. Make asking part of every new customer's bill.
Judging after one week
Some channels build slowly. Give it the full period unless it's clearly bringing nobody.
Buying customers with deep discounts
Heavy discounts bring people who leave when the offer ends. Measure who comes back at full price.
Go deeper
- Read nextPlan an opening offer →The launch version: a short offer to bring the first customers, before you have numbers.
- Business playbookGetting customers: tuition centre →Channels a tuition centre can test before each admission season: schools, societies and referrals.
- Business playbookGetting customers: gym →How a gym measures trial passes and local promotions against memberships that renew.
The playbooks are written for one business each. Start with the full playbook or find yours.
Questions people ask
Which marketing channel works for a local shop in India?
It depends on who your customer is and where they spend time. Societies and schools suit some businesses; local search ads or a delivery app suit others. That's why you test one at a time with your own numbers instead of copying another shop.
Should I pay for ads on Google or social media?
Only after your free profile and reviews are in order, and only as a test with a fixed budget and a way to count results. Start with a small radius around the shop and a short period.
Is a referral offer a channel worth testing?
Often, because it uses customers who already like you. Offer something small to both the regular and the friend, and count it with a code. Your customer list is where you'd announce it.