Skip to content

Performance

What a slow store actually costs you in Pakistan

Most Pakistani ecommerce traffic arrives on a mid-range Android over congested 4G. Here is the arithmetic on what every extra second removes from your revenue.

By TezCommerce · · 6 min read

There is a specific kind of conversation we have with store owners who are spending real money on Meta ads and cannot work out why the orders are not there. The creative is fine. The targeting is fine. The product sells perfectly well over WhatsApp. The website is the problem, and the owner has never seen it behave the way their customers see it behave.

You are not your customer's device

You look at your store on a laptop, on office wifi, with the page already cached. Your customer taps an Instagram ad on a three-year-old Android, on a congested cell in Nazimabad at 9pm, with fourteen other apps holding memory. Those are not the same website. The second one is frequently four or five times slower, and it is the only one that matters commercially.

This is why we test builds on real low-end devices with network throttling rather than trusting a desktop Lighthouse score. A store can report 98 on a developer's machine and still take six seconds to become interactive for the person you paid Rs. 18 to bring there.

The arithmetic

Work it through with round numbers. Say you spend Rs. 100,000 a month on Meta, and your cost per click lands at about Rs. 18. That buys roughly 5,500 sessions. At a 2% conversion rate and an average order value of Rs. 2,500, that is 110 orders and Rs. 275,000 in revenue.

Now make the site slow. Studies disagree on the exact figure, but they agree on the direction: every extra second of loading costs conversions, and bounce rates rise sharply past three seconds on mobile. Suppose, for illustration, that the store at six seconds converts at a quarter of the rate it manages at two. That is 27 orders instead of 110, on identical spend.

Where the seconds actually go

In our audits of Pakistani stores, the same handful of causes account for most of the damage, roughly in this order:

  1. Uncompressed product photography. A 4MB hero image straight from a phone camera is the single most common cause. Served as WebP at the right dimensions it is often under 150KB with no visible difference.
  2. Template bloat. A purchased theme loading six carousel libraries, three font families and an animation framework for one fade effect.
  3. Third-party scripts. Chat widgets, review apps, popup tools and three separate analytics tags, each blocking the main thread.
  4. No caching or CDN. The page rebuilt on every request from a server in another continent.
  5. Web fonts loading render-blocking, so the text is invisible until the font arrives.

What good looks like

The targets we build to, measured on a throttled mobile connection rather than a desktop: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1, and a total initial payload under 500KB. Those are not aspirational: they are achievable on a Rs. 49,000 build, and they are the difference between paying for traffic and paying for orders.

How to check yours in five minutes

  • Open PageSpeed Insights and run your product page, not your homepage. Read the mobile score, ignore the desktop one.
  • Open your store on your own phone, on mobile data rather than wifi, after clearing the cache.
  • Count the seconds until you can actually tap Add to Cart. That number is your real conversion ceiling.

If that number is above three, the cheapest growth available to you this quarter is not a bigger ad budget. It is fixing the page the budget points at.

Start a project

Let's work together.

Tell us where you want the business to be in a year. We will tell you how we would get you there, what it would cost and how long it would take, in writing.