A customer taps your link on a phone, on mobile data, in a shop with one bar of signal. If the screen is still white after three seconds, they are gone — back to the search results, tapping the next business. You never see them, and your analytics records a bounce you will probably blame on the traffic quality.
Speed is the cheapest conversion improvement available to a small business, and the most neglected. Here is what actually causes the problem, and what is worth fixing.
What actually slows down an Indian small business website
In our audits, the same five causes appear again and again — and none of them is "the internet is slow".
- Uncompressed images. A single photograph straight from a phone camera can be 6 MB. Eight of them on a page is a 48 MB download. Compressed and served in a modern format, the same page is under 1 MB and looks identical.
- Page builders and plugin stacks. A typical WordPress site with a page builder, a slider, a form plugin, a chat widget and three font plugins loads dozens of scripts before it renders a single word.
- Render-blocking CSS and JavaScript. The browser cannot draw the page until it has fetched and parsed certain files. If those files are large or slow to arrive, the visitor stares at white.
- Web fonts loaded badly. Four weights of three different fonts, each requested from a different server, each blocking text from appearing.
- Cheap shared hosting in the wrong region. A server in the United States serving a customer in Bihar adds several hundred milliseconds to every single request, before a byte of content is transferred.
The three numbers that matter
Google measures speed with three Core Web Vitals. You do not need to memorise them — but you should know what each one means, because each has a different fix.
| Metric | What it measures | Good target | Usual cause |
|---|---|---|---|
| LCP | How long until the main content is visible | under 2.5 s | Big hero image, slow server |
| INP | How quickly the page responds to a tap | under 200 ms | Heavy JavaScript, too many plugins |
| CLS | How much the layout jumps while loading | under 0.1 | Images without dimensions, late-loading banners |
If you only track one, track LCP. It is the one a customer feels as "this site is slow".
How to measure it honestly
- Open PageSpeed Insights and enter your URL. Read the mobile tab, never the desktop one. Your customers are on phones.
- Open your site on an actual mid-range Android phone, on mobile data, not office wifi. Count the seconds until you can read the headline.
- Check Search Console → Core Web Vitals for the field data, which is real-user measurement rather than a lab test.
A lab score of 90 with a real-user LCP of five seconds means the lab test is being fooled. Trust the field data.
The fixes that are actually worth paying for
1. Compress and convert every image
Resize images to the dimensions they are displayed at, convert to WebP or AVIF, and set width and height attributes so the layout does not jump. This one change usually cuts page weight by more than half and costs almost nothing.
2. Stop loading what nobody sees
Chat widgets, sliders, animation libraries and tracking scripts that are not needed on the first screen should load after the page is usable, or not at all. Audit every script and ask what it does for the customer. If the answer is "nothing", remove it.
3. Serve fonts properly
Two font families maximum, the weights you actually use, self-hosted or preconnected, with font-display: swap so text appears immediately in a fallback face.
4. Host close to your customers
For an Indian audience, host in India or Singapore, on a CDN that has an Indian edge. This is the difference between a 40 ms and a 400 ms first byte.
5. Fix the server, not just the page
If your host takes two seconds to respond before sending any HTML, no amount of image compression will save you. Test the time to first byte. Above 600 ms, change hosts.
If a fix cannot be explained in one sentence and measured before and after, it is not a performance fix. It is a redesign in disguise.
What speed is worth to a business
We will not quote a percentage from a study conducted in another country. The honest version is simpler: look at your own analytics. Take the number of visitors per month, and assume a modest share of them leave because the page is slow. Multiply by your average enquiry value. That number is what you are paying every month for a page that takes five seconds instead of two.
For most local businesses that calculation produces a figure far larger than the cost of fixing it — which is why speed is usually the first thing we work on, before content, before links, before anything fashionable.
Fix or rebuild?
Fix, if the site is structurally sound: clean HTML, a reasonable theme, few plugins. Rebuild, if the site is a page-builder stack with thirty plugins, no staging environment and a theme nobody has updated since 2021. In that case, every fix is temporary and the next update undoes it.
We usually know within an hour of opening a site which of the two it is, and we say so plainly — including when the answer is "do not spend money on this, fix these four things and stop".
A five-minute self-check
- PageSpeed mobile score above 70?
- LCP under 2.5 seconds on a real phone?
- Total page weight under 1.5 MB?
- Fewer than 15 scripts on the page?
- Images in WebP or AVIF, with width and height set?
- Time to first byte under 600 ms?
Four or more "no" answers means speed is your cheapest available win. Book the free audit and we will show you exactly which files are responsible.