Most D2C brands don't fail at ads. They fail at the product decision, months before the first ad ever runs.
The pattern is familiar. A founder falls for an idea, orders a first batch, builds a beautiful store, and starts spending. The ads get clicks. The orders don't come, or they come once and never again. By the time the problem is obvious, most of the capital is sitting in a godown as stock.
None of that is an ad problem, which is exactly why no amount of creative testing fixes it. The product was chosen on instinct, and instinct is the most expensive research method in D2C.
This is the framework we'd use instead. Five steps, each cheaper than the next mistake, and a scorecard at the end so you can rank your ideas honestly.
The framework at a glance
- 01
Start with a repeat problem, not a product.
Write the problem in one line before you name anything.
- 02
Pick boring.
Proven demand, with a flaw you can obviously fix.
- 03
Check demand in ten minutes.
Google Trends, marketplaces, Meta Ad Library.
- 04
Run the repeat and margin maths.
If the numbers don't work on paper, they won't work in the ad account.
- 05
Prove it with money.
Get someone to pay, or commit to pay, before you order inventory.
Step 1: Start with a repeat problem, not a product
A good first product has to pass three tests at once.
Test 1
You understand the problem it solves
Ideally because you've had it.
Test 2
You can get in front of the people who have it
Without spending a fortune.
Test 3
There's enough money left in each sale
To pay for delivery and ads.
So before you name a product, write the problem in one line.
Not this
- "A premium face serum"
- "Cotton loungewear"
But this
- "Women in their late twenties with oily skin can't find a vitamin C serum that doesn't break them out."
- "Working women want loungewear they can wear on a video call."
The one-line version does two jobs. It tells you whether the problem is real and recurring, and it becomes the opening line of every ad you'll run later. If you can't write it, you don't have a product idea yet. You have a category.
Many of India's best-known D2C brands started exactly here. Mamaearth's founders were importing toxin-free baby products from the US for their own child and found other parents doing the same. They launched with six baby-care products, not a range. The Whole Truth started from its founder's own frustration with "healthy" snacks full of hidden sugar, and launched one product: a protein bar.
Neither started with "what's trending." Both started with a problem they kept running into.
Step 2: Pick boring
First-time founders love new categories, because new feels like an opportunity. It usually isn't.
A new product means nobody is searching for it. You aren't selling a product any more, you're creating a habit, and creating a habit is the most expensive thing an ad budget can be asked to do.
So look for the opposite: a product people already buy in large numbers, where the current options keep disappointing them. Maybe the fit is off, the colours are dull, the sizes stop at L, or the fragrance gives people headaches. A busy category full of visibly unhappy customers is a far better bet than an empty one nobody is shopping in.
How to find the flaw: review mining
The research is free and sitting in public.
- 01
Pick the marketplace.
Search your category on Amazon, Flipkart, Nykaa or Myntra, whichever your customer actually uses.
- 02
Open the top ten sellers.
- 03
Ignore the five-star reviews.
Read the one-, two- and three-star ones, at least 30 per listing.
- 04
Log every complaint.
A simple sheet: one column for the complaint, one for how many times you've seen it.
- 05
Sort by count.
What you're looking for is the same complaint, repeated across several brands. "Runs small." "Fades after two washes." "Too sticky in humid weather."
If the top complaints are things you can fix, and fix visibly, you have a reason to exist. If the complaints are random, or the top sellers are genuinely good, find another category.
Step 3: The 10-minute demand check
Asking friends is not validation. They'll say it's nice, because that's what friends do. These three checks use real behaviour instead, and together they take about ten minutes.
Check 1 · Google Trends
Is anyone curious?
Five years of search interest, not one month.
Check 2 · Marketplaces
Is anyone buying?
Rating counts on the top five sellers.
Check 3 · Meta Ad Library
Is anyone profitable?
How long other brands keep paying for the same ad.
Check 1: Google Trends — is anyone curious?
Search your product on Google Trends, set the country to India and the range to five years, not one month.
What the five-year line tells you
Steady or slowly rising line
Good
Real, durable interest.
Flat near zero
Wrong term, or no market
Try the words your customer would actually type.
A sharp spike that collapsed
Too late
A trend you've already missed.
Scroll down to "interest by subregion". The states searching most are where your first ads should go.
Check 2: Marketplaces — is anyone buying?
Google tells you people are curious. Amazon and Flipkart tell you people are paying.
Search the product and look at the top sellers. If the top five each carry thousands of ratings, people are buying this every day. If the best seller has 40 reviews, the demand you saw on Google hasn't turned into purchases yet.
Check one more thing while you're there: the price band. Note where most of the top sellers sit. That's the price your customer already expects, and you'll need a clear reason to sit above it.
Check 3: Meta Ad Library — is anyone profitable?
This is the most underused free tool in D2C. Open the Meta Ad Library, set the country to India, and search your product.
Then look at the start date on each ad. If several unrelated brands have been running ads for the same product for 30, 60 or 90 days, that is the strongest demand signal you can get for free. Nobody keeps paying for an ad that isn't making money. Someone else has already proven the demand, with their own budget.
While you're there, save their best-running ads. They tell you which angles, offers and price points the market already responds to.
Step 4: Run the repeat and margin maths
Demand tells you people will buy. It doesn't tell you whether you'll make money when they do. That's decided by two numbers you should know before you pick a product: how often the customer comes back, and how much margin survives each order.
The first order usually loses money
Acquiring a new customer in Indian D2C now costs somewhere around ₹800 to ₹1,200. Put that against a typical first order.
Say the product sells for ₹1,000. After product cost, packaging, shipping, gateway fees and returns, you might keep ₹500. You've spent about ₹1,000 to make ₹500.
₹800–1,200
Typical cost to acquire a new customer in Indian D2C
₹1,000
Selling price of the first order
~₹500
What you keep after product, packaging, shipping, gateway and returns
That isn't a disaster. It's normal. But it means the business only works if that customer comes back, because the second and third orders arrive without the acquisition cost.
So ask one simple question of every product idea: will they run out of it? Face wash runs out in a month. Protein, coffee and supplements run out. A mattress is bought once in several years. Repeat isn't everything, but if your product is a one-time purchase, you need a much higher price or a much cheaper way to find customers. Know which one you're building before you start.
Gross margin decides more than you think
A study of 34 Indian D2C brands, built from their RoC and DRHP filings, found gross margin was the single strongest predictor of which brands reached profitability.
Gross margin vs. profitability, across 34 Indian D2C brands
Above ~70%
The profitable leaders
Where the profitable leaders mostly sat.
Below ~60%
A struggle
Brands struggled to make the numbers work.
Below 45%
No one at scale
No Indian D2C brand had become profitable at scale without owning its manufacturing.
That's why so much D2C money sits in skincare and beauty, where 65–72% gross margins are normal, and so little in fresh food or electronics.
The same ₹1,000 order in two categories
| Per order | High-margin category | Thin-margin category |
|---|---|---|
| Selling price | ₹1,000 | ₹1,000 |
| Product + packaging | ₹300 (70% margin) | ₹550 (45% margin) |
| Shipping + fulfilment | ₹100 | ₹100 |
| Payment gateway (~2.5%) | ₹25 | ₹25 |
| Returns / RTO, spread per order | ₹30 | ₹53 |
| Left before ads | ₹545 | ₹272 |
| Break-even ROAS | 1.83x | 3.68x |
How break-even ROAS is worked out
High-margin: 1 ÷ (₹545 ÷ ₹1,000) = 1.83x. Thin-margin: 1 ÷ (₹272 ÷ ₹1,000) = 3.68x.
Same price, same customer, same ads. One product needs a 1.8x ROAS to break even, the other needs 3.7x. The second founder will spend every month thinking they have an ad problem. They don't. They picked the category.
Do this for your own idea before you place an order. If what's left before ads won't cover the cost of a customer within two or three orders, change the product, the price, or the bundle, not the ad account later.
Step 5: Prove it with money before you order inventory
The first four steps are research. This one is evidence. Until someone has paid, or at least committed to pay, you have a hypothesis.
Work up this ladder, cheapest rung first, and stop when you have a clear answer:
The proof ladder: cheapest rung first
- 1
Ask about the past, not the future
Free
"When did you last buy a face serum, where from, and what annoyed you about it?" gets you real answers. "Would you buy this?" gets you politeness. People describe their past accurately and their future generously.
- 2
Run an ad test before stock
A few thousand rupees per variant
Use a sample or good mock-ups and run five to ten creative variants on Meta. Watch click-through and add-to-cart rate. You aren't trying to be profitable yet, you're trying to learn whether strangers care.
- 3
Send the traffic somewhere real
Pass mark: 2–5%
One page with a clear promise, real or mocked-up photos, a real price, and a "Reserve yours" or "Notify me" button. Cold traffic usually converts at 2–5%. Under about 2–3%, rework the concept, not just the ad.
- 4
The Indian shortcut: ads straight to WhatsApp
Small budget
A small budget sending people to a WhatsApp chat shows you, in their own words, who asks for the price, the size and the delivery date.
- 5
Take paid pre-orders
Real money
Presell on Instagram or your Shopify store with an honest delivery date. Global crowdfunding platforms are hard to use from India, so a plain pre-order on your own Shopify store is usually the easier route.
- 6
Sell a small batch where buyers already are
First stock
Wakefit's founders sold a few hundred mattresses on Amazon before building their own brand site, so they knew their real costs before they scaled.
For a first batch in India you rarely need a huge list. A few hundred to a thousand people who have genuinely raised their hand on WhatsApp or Instagram is usually enough. You don't need a viral launch. You need proof that strangers will pay.
The product scorecard
Score each product idea against these seven questions. Be strict: give the 2 only when you have the evidence in front of you, not when you believe it.
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 1. Problem: can you write a real, recurring problem in one line? | No clear problem | A problem, but occasional | Clear, frequent, and you've lived it |
| 2. Search demand: what does Google Trends show over 5 years? | Flat or a dead spike | Low but steady | Steady or rising |
| 3. Buying demand: how many ratings do the top 5 marketplace sellers have? | Under 100 each | Hundreds | Thousands |
| 4. Fixable flaw: do reviews show a repeated complaint you can fix? | No pattern | A pattern, hard to fix | A clear pattern you can visibly fix |
| 5. Paid proof: are unrelated brands running ads for 30+ days? | None | One or two | Three or more |
| 6. Repeat: does the customer run out and re-buy? | One-time purchase | Every 6–12 months | Within about 90 days |
| 7. Margin: what gross margin can you realistically hit? | Under 45% | 45–60% | 60%+ |
Total out of 14
11–14
Go
Move to Step 5 if you haven't already, and order a small first batch.
8–10
Fix first
Find the weak rows. A thin margin can often be fixed with a bundle or a higher price; low search demand usually can't.
Under 8
Drop it
Keep the research. It will help with the next idea.
Run three or four ideas through it side by side. The point isn't to find a perfect score. It's to stop the idea you're emotionally attached to from beating the idea the market actually wants.
What to do this week
Here's what to do:
- Write your idea as a one-line problem. If you have several ideas, write one for each.
- Spend an hour mining the 1–3 star reviews of the top ten sellers in your category. Count the repeated complaints.
- Run the 10-minute demand check on each idea: Trends over five years, marketplace rating counts, Meta Ad Library start dates.
- Fill in the margin table with your own quotes for product, packaging and shipping. Work out your break-even ROAS.
- Score every idea out of 14 and drop anything under 8.
- For the winner, spend a small amount on an ad test or a pre-order page before you place a production order.
If an idea survives all six, you've done more research than most brands that go on to spend lakhs on ads. And if it doesn't, the research cost you a few evenings instead of your first batch of stock.
Picking the product is the first big decision, and the one every later decision depends on. Your margin, your ROAS and your repeat rate are mostly decided here, long before the first ad goes live.
Already past this stage and doing ₹10L+ a month? Then your question is no longer what to sell, it's how to scale it profitably. That's the work we do, and the free audit below is where it starts.