Three Numbers Decide Whether Your Store Makes Money
Acquisition costs are not going back down. The stores that adapted did it by improving what happens after the click, not by finding cheaper clicks.
By Ads Ninza · Published 26 August 2026 · 12 min read
There is no cheaper traffic
Rising acquisition costs have permanently changed e-commerce economics. A store that worked at earlier click prices does not automatically work now, and the businesses that adapted did not do it by finding a cheaper channel. They did it by improving what happens after the click.
Three numbers decide the outcome, and they multiply rather than add. What share of visitors buy. What they spend when they do. How often they come back. A modest improvement in each transforms the economics far more than any bid strategy adjustment, because the gains compound against each other.
The fourth number sits underneath all of them: gross margin after fulfilment. A store can be beautiful, convert well, and lose money on every order. Which is why we start from the margin sheet rather than from the design.
Product pages are where the decision happens
Most store optimisation attention goes to the home page. Most purchase decisions happen on the product page, and most product pages answer fewer of the buyer questions than they should.
The test is simple: list every question a buyer needs answered before purchasing. Sizing and fit. Materials and construction. Delivery timing and cost. Returns, plainly stated rather than buried. How this compares to the cheaper option. Whether it suits their specific situation. Then check how many the page actually answers.
In most stores it is about half, with the rest either absent or hidden in a policy page nobody opens. Every unanswered question is a reason to leave and check a competitor, and you paid for that visit.
Photography is the other half of this, and we will be direct that it is usually the constraint rather than the copy. Product photography is one of the highest-impact assets in e-commerce and the one most often under-invested in.
Average order value, the lever nobody works
Raising average order value increases revenue per customer without increasing acquisition cost at all. It is the cheapest growth available and most stores do almost nothing here.
Bundles that genuinely belong together. Not arbitrary pairings to hit a price point, but combinations a buyer would plausibly want, priced so the bundle is obviously better value than the parts.
Free shipping thresholds set just above current AOV. If your average order is 1,400 rupees, a threshold at 1,500 moves a meaningful share of orders upward. Set at 3,000 it is ignored.
Cross-sells at the right moment, and the word "considered" matters. Irrelevant recommendations add friction and reduce conversion, which costs more than the upsell gains. A cross-sell that does not make sense is worse than none.
Post-purchase offers. The moment immediately after someone has bought is the highest-intent moment you will ever have with them, and most stores use it to show a generic thank-you page.
Repeat purchase is where the margin actually lives
At current acquisition costs, the first order frequently breaks even or close to it. Profit comes from the second and third, which makes retention infrastructure a growth investment rather than a nice-to-have that gets deferred.
Practically that means email and WhatsApp flows built as part of the store rather than added later: post-purchase sequences, replenishment timing for consumables, win-back for lapsed buyers, and a reason to return that is not always a discount. Discounting your way to repeat purchase trains customers to wait for the discount.
This also changes how you read your ad reporting. If your reporting does not separate new from returning customers, returning buyers subsidise your acquisition numbers and you systematically overestimate how well you are acquiring. Splitting them is uncomfortable the first month and necessary.
Cash on delivery, honestly
In Indian e-commerce, COD raises conversion rate and raises return rates together, and the second effect is frequently larger than the celebration around the first.
A failed delivery costs you shipping both directions, handling, and the working capital tied up in the interim. A store reporting excellent conversion and poor profitability is often a COD story rather than a marketing one.
What helps: partial prepayment options with a small incentive, address verification before dispatch, an order confirmation flow on WhatsApp that catches wrong numbers and changed minds early, and reporting contribution margin per order rather than revenue. None of that removes the effect. It makes it visible and manageable, which is the realistic goal.
Frequently asked questions
Should I use Shopify or something else?
Shopify suits most businesses well because the ecosystem, payment handling and reliability are strong and the platform stays out of the way. It is less suitable for highly unusual fulfilment logic or very large complex catalogues. We will tell you if your requirements point elsewhere rather than building on Shopify by default.
What is a good ROAS for e-commerce?
There is no universal figure, and treating ROAS as the goal is itself the problem. A 3x return on a thirty per cent margin product is a loss once fulfilment is counted. Contribution margin after cost of goods, shipping, payment fees and returns is the number that tells you whether to scale.
How do we increase average order value?
Bundles that genuinely belong together, quantity incentives, free-shipping thresholds set just above your current average, considered cross-sells at the right moment, and post-purchase offers. The word considered is doing real work there: irrelevant recommendations reduce conversion and cost more than they add.
Does site speed affect sales?
Yes, measurably, and mobile matters most. The drop is steepest in the first few seconds. The usual culprits are oversized images and an accumulation of apps nobody audited. Speed also influences your ad costs indirectly through landing page experience signals.
How long does a Shopify build take?
Four to eight weeks from research to launch for a standard store, depending on catalogue size, integrations and how ready your product content is. Migrations take longer because data mapping and URL redirects need care. Photography and copy readiness on your side is usually the critical path.
Can you improve our existing store instead of rebuilding?
Often, and it is usually the better first step. We audit conversion rate, speed, product pages and checkout, then fix in priority order. A rebuild makes sense when the theme is fundamentally limiting or the architecture is wrong.
Do you handle product photography?
Not in-house, and we will not pretend otherwise. We specify exactly what images each product page needs and work with your photographer, or recommend one. It is too important to fake competence at.
How important are Google Shopping and Meta catalogue feeds?
More than most stores assume. Feed quality directly determines Shopping and Advantage+ performance, because titles, attributes, GTINs and category mapping decide what your products compete for. Most feeds we audit leave performance on the table through incomplete data rather than through bidding.
How should we handle cash on delivery?
Deliberately. Offer partial prepayment with a small incentive, verify addresses before dispatch, confirm orders on WhatsApp to catch errors early, and report contribution margin per order rather than revenue. COD raises conversion and return rates together, and the margin model has to account for both.
Do you run the ads as well as building the store?
Yes, and it works better when one team owns both. When store and media are managed separately, conversion problems and traffic problems get blamed on each other indefinitely while nothing improves. Services start at INR 35,000 per month in India and $500 outside it.