The ecommerce CRO checklist we run before adding budget

Before we agree to scale anyone's ad spend, we run the store through this list. It takes a day. It has saved more accounts than any bidding change we have ever made, because doubling traffic into a store converting at 0.7% just doubles the cost of the same problem.

Across the Indian D2C stores we work with, sitewide conversion rates typically land between 1.1% and 2.6%, with mobile running roughly a third below desktop. If you are under 1%, the problem is almost never your targeting.

Speed, measured on the phone your customer owns

Test on a mid-range Android over 4G, not on your MacBook on studio wifi. In our audits this single test explains more lost revenue than anything else on the list.

• Largest contentful paint under 2.5 seconds. Above four, expect to lose a third of arrivals before the page paints.
• Compress every image to WebP or AVIF. Hero images above 200KB on mobile are a choice, not a constraint.
• Remove abandoned apps and pixels. We regularly find eight to fourteen scripts on Shopify stores, half from tools nobody has opened in a year.
• Lazy-load below the fold, never above it.

The product page does the selling

Most Indian D2C product pages are a gallery and a buy button. That is a catalogue, not a sales page. What we look for:

1. Price and offer visible without scrolling on a 6.1-inch screen.
2. A delivery estimate by pincode, on the page, before checkout. "Delivers to 560038 by Tuesday" converts better than any trust badge ever printed.
3. Six to nine images minimum, including scale reference, texture macro, in-use, and packaging.
4. Reviews with photos, on the page, not behind a tab. Tabs hide the single most persuasive asset you own.
5. A size chart in Indian measurements with a fit note, for apparel. Sizing anxiety is the top apparel drop-off we see.
6. Answers to objections in the copy. Is it safe for sensitive skin. Does it work on Indian hair. What happens if it does not fit.

Write the ingredient or material story properly. Vague claims read as evasive, and Indian shoppers have been trained by Nykaa and Myntra listings to expect specifics.

The COD question

COD is not free. It is a conversion subsidy you pay for in RTO. Typical ranges from stores we work with:

Removing COD usually drops orders sharply, so do not do it. Do this instead.

• Offer a small prepaid incentive, 5% or free shipping. Cheaper than one RTO.
• Add an OTP or WhatsApp confirmation step on COD orders above a threshold.
• Suppress COD for repeat-RTO pincodes. Every store has a list. Pull it.
• Send a WhatsApp confirmation before dispatch. This alone cuts RTO meaningfully in the accounts where we have implemented it.

Checkout, where the money leaks

• Guest checkout, always. Forced account creation is pure loss.
• Ask for the minimum. Every optional field costs you conversions.
• Auto-fill city and state from pincode. Nobody wants to type Karnataka.
• Show the full landed price early. Surprise shipping at the final step is the most common abandonment cause we can identify from session recordings.
• UPI first in the payment list, above cards. It is how India pays.
• Keep the coupon field small or hidden behind a link. A prominent empty coupon box sends people to Google to hunt for a code, and many do not come back.

Watch twenty session recordings of people who abandoned checkout. You will find your two biggest problems in an afternoon, and neither will be the one your team was arguing about.

Trust signals that are actually read

Return policy in plain language, linked near the buy button. A real address and phone number in the footer, ideally a working one. Order tracking that exists. Genuine review counts, because inflated ones are obvious and cost more credibility than they buy.

Test with discipline

Most stores do not have the traffic for meaningful A/B testing. If you are under roughly 40,000 monthly sessions, do not split-test button colours, you will never reach significance. Make large, confident changes and read the before-and-after over four weeks, holding traffic mix roughly constant.

Reserve real testing for the few pages with volume: the top-selling product page, the cart, and the homepage. Test one thing at a time. Give each test at least two full weeks so weekday and weekend behaviour both land in the data.

Typical uplift ranges we see when the fixes above are implemented properly:

The order of operations

Fix speed. Then the product page. Then checkout. Then COD economics. Only then increase spend. Doing it the other way around is how brands end up with a large ad bill and a store that was never ready for the traffic.

If you want us to run this audit on your store, we do it as part of a free strategy call from our Indiranagar studio. Bring analytics access and your RTO numbers, and we will show you where the leak is before anyone talks about media budgets.