Kodakod Team
In e-commerce, the most expensive mistakes are made before the site ever goes live. Here are 8 common mistakes and their fixes, drawn from the projects where we keep correcting the same problems.
In short: The costliest mistakes when launching an e-commerce site are a long checkout with mandatory registration, an interface squeezed onto mobile as an afterthought, product descriptions copied from the supplier, uncompressed images, missing trust and legal information, unmeasured conversion, stock that is not connected to the site, and no plan for life after launch. Every one of them goes back to a decision made during setup, so the cheapest fix is to check these eight points one by one before the site goes live.
None of the mistakes in this article came out of a book. The first six are real cases from e-commerce projects we inherited or were called in to fix; the last two are the topics most easily put off as "we'll deal with it later" during setup. What they have in common is this: not one was noticed on launch day. All of them surfaced once traffic started growing.
Below we describe each case as we found it — the symptom, the underlying cause, and what we did.
Case: "Plenty of visitors, no sales"
Symptom. A clothing brand spending seriously on ads was happy with visitor numbers but had almost no sales.
What we found. Checkout ran to five steps, and step one was mandatory registration. Users were forced to create an account after adding an item to the cart.
What we did. Enabled guest checkout, cut it to two steps, and moved shipping costs from the cart to the product page. A surprise shipping charge is the quietest cause of abandoned carts.
Case: The revenue lost on mobile
Symptom. Three quarters of traffic came from mobile, yet most sales came from desktop. The company had read this as "mobile users are just browsing."
What we found. The site had been designed on desktop and squeezed into mobile. On a phone the filter menu fell below the fold, and the "add to cart" button disappeared when the keyboard opened.
What we did. Rebuilt the interface mobile-first. That was the interpretation, but here is the reality: mobile users were not browsing, they were unable to buy.
Case: The 400 products Google never saw
Symptom. A wide catalog, but organic traffic close to zero.
What we found. Every product description had been copied from the supplier's catalog. The identical text sat on dozens of other resellers' sites.
What we did. Rewrote the descriptions for the 40 best sellers from scratch — who it suits, what it pairs with, how it fits. A search engine needs a reason to favour you; the same sentences are not a reason.
Case: The homepage that took 12 seconds
Symptom. Most people arriving from ads bounced before seeing the page.
What we found. The campaign images on the homepage had been uploaded exactly as the designer delivered them, uncompressed. A single image ran to several megabytes.
What we did. Converted the images to modern formats, served them at responsive sizes, and deferred anything below the fold. The painful part: a click fee had been paid for every one of those visitors.
Case: The site nobody trusted
Symptom. Good products, competitive pricing, but low conversion.
What we found. No contact address on the site, a one-line return policy, and no terms of sale. For a first-time visitor there was not a single trust signal.
What we did. Put clear contact details, a plain return policy and genuine customer reviews where they could be seen. These are as much a legal requirement as they are the signals someone looks for before typing card details into a site they do not know.
Case: Not knowing what was broken
Symptom. The company knew sales were low but not where they were losing people.
What we found. No conversion tracking at all. Only total visitor count was being measured.
What we did. Started measuring add-to-cart, checkout start and purchase separately. Within the first week it became clear which step was losing them. You cannot fix what you do not measure — but worse, you cannot tell whether you fixed it.
Mistake: Not connecting stock and orders to the site
Symptom. A product shown as "in stock" on the site is not in the warehouse; once the order arrives you have to call the customer to ask them to wait, or cancel. If you also sell on marketplaces, the same item gets sold through two channels at once.
Cause. Stock, prices and orders are updated by hand; accounting, warehouse and shipping do not talk to the site. That works at low volume, but every new order means one more manual step, and the margin for error grows with every order.
Fix. Plan at the setup stage for stock to be managed in a single source (an ERP, accounting or inventory system) and pushed from there to the site and marketplaces. Set up the shipping-carrier integration too, so the tracking number reaches the customer automatically. Thinking about integration from the start is far cheaper than changing the infrastructure later.
Mistake: Not planning for after launch
Symptom. The site launches and the first months go smoothly. Then a plugin update breaks the checkout page, the SSL certificate expires or the server has a problem, and there is no recent backup.
Cause. An e-commerce site is treated like a brochure that is built once and left alone. In reality the payment integration, plugins, theme and server software are updated constantly, and order history and customer data are assets that need protecting.
Fix. When you collect proposals, ask about life after launch too: who will apply updates and how often; are backups taken automatically, and has restoring from one ever been tested; who do you call when the site is down? Testing updates in a staging environment before applying them to the live site keeps you from finding the checkout broken one morning.
Pre-launch checklist
Each of the eight mistakes above could have been caught with a few hours of checking before the site opened. Go through this list before your launch date:
- Checkout flow: Is guest checkout enabled, and has the offer to create an account been moved to after the order? According to data compiled by Baymard Institute, setting aside people who were just browsing, 18% of shoppers abandon a cart because the site wanted them to create an account.
- Order total: Are the shipping cost and the order total visible before the cart? In the same data, the most common reason for abandonment, at 40%, is that extra costs such as shipping, tax and fees were too high.
- Mobile: Have finding a product, filtering, adding to cart and checking out been tried end to end on a real phone, and is the “add to cart” button still visible when the keyboard is open?
- Product descriptions: Are the descriptions of the products you expect to sell most original text rather than the supplier's copy?
- Images: Are images compressed, in a modern format and served at a size that fits the screen? According to web.dev, sending desktop-sized images to mobile devices can use 2–4 times more data than needed. The main image visible on the first screen, however, should not be lazy-loaded; deferring is only for images further down the page. The speed metrics are covered in our article on Core Web Vitals.
- Trust and legal information: Are a full address and contact details, return terms, the pre-contract information and the distance sales contract in place? Article 5 of Turkey's Distance Contracts Regulation (Mesafeli Sözleşmeler Yönetmeliği) requires the consumer to be informed, before the contract, of the seller's full address and contact details, the total price including taxes, extra costs such as shipping, and the conditions of the right of withdrawal. If extra costs are not disclosed, the consumer is not liable to pay them, and the burden of proving that the information was given lies with the seller.
- Measurement: Are add-to-cart, checkout start and purchase measured separately? Use a test order to confirm that every step is recorded. We explain how to read funnel data in our article on conversion rate.
- Stock and orders: Is stock managed from a single source, and do the site and the marketplaces see the same stock? Check that a test order is deducted from stock and that the shipping tracking number reaches the customer automatically.
- After launch: Is it written down who applies updates, that automatic backups are taken and restoring from one has been tried at least once, that the SSL certificate renews automatically, and whom to call when the site is down?
If you are still choosing a platform, we walk through the order of these decisions step by step in our guide to building an e-commerce site.
What these mistakes share
In none of the eight was the problem "not advertising enough." In all of them there was infrastructure that could not carry the traffic or the orders already arriving. And in all of them the root cause traced back to a decision made while the site was being built.
That is why the most valuable conversation in an e-commerce project is the first one. Building a system correctly from the start is almost always cheaper than correcting it later. If you are at the setup stage and one of these mistakes sounded familiar, let's talk: in our e-commerce solutions we plan these decisions together with the site from the start.
Frequently asked questions
The most common mistakes are a long checkout with mandatory registration, an interface designed for desktop and squeezed onto mobile, product descriptions copied from the supplier, heavy images uploaded without compression, missing contact and return information, and conversion tracking that was never set up. Add to these stock and orders that are not connected to the site, and no plan for maintenance and backups after launch. What they share is that the root cause goes back to a decision made while the site was being built.
When the same description appears word for word on dozens of resellers' sites, the search engine has no reason to favor your page. That is why organic traffic can stay very low even with a large catalog. The fix is to write original descriptions, starting with the best sellers, that explain who the product suits, what it pairs with and how it fits.
This often means mobile users are not just browsing but are unable to buy on their phones. On sites designed for desktop and squeezed onto mobile, the filter menu can fall below the fold or the add-to-cart button can disappear when the keyboard opens. The fix is to build the interface mobile-first from the start and test the purchase steps on a real phone.
These are both a legal requirement and the trust signals a first-time visitor looks for; a distance sales contract is the contract Turkish consumer law requires for online sales to consumers. A user about to enter card details on an unfamiliar site wants to see a clear contact address, a plain return policy and genuine customer reviews. When these are missing, conversion can stay low even if the products and prices are good.
Because most of them only become visible once traffic starts to grow; with few visitors at launch, everything looks fine. The way to catch them early is to measure add-to-cart, checkout start and purchase separately from day one, so it quickly becomes clear at which step people are lost. Building a system correctly from the start is almost always cheaper than fixing it later.
