Fix the theme or build a store from scratch?

Founders ask for a new store when they are tired. That is a mood, not a diagnosis. A store from scratch is the right call when identity, catalog, and checkout cannot be untangled. A theme fix is the right call when the platform fits and the faults are weight, contrast, and apps.

x8x does both. We do not sell templates. We either repair the theme you run on Salla, Zid, Shopify, or WooCommerce, or we build a full store on the platform that matches your volume. The scan exists so you do not buy a rebuild you do not need, and so you do not patch a theme that will never RTL.

Decide after a measurement, not after a mood

Open the live URL you advertise. Scan it. Read LCP, the screenshot, HTTPS, and the button row. If the only red is a hero image and a chat app, you are in fix territory. If the screenshot is an English layout with clipped Arabic, three pay buttons, and a cart that jumps, you are in build territory.

Mood says "burn it." The report says which files and which path. Keep the mood out of the purchase order.

What "from scratch" means here

It does not mean a theme pack with your logo dropped in. It means we design the catalog structure, the product template, the RTL checkout, shipping and pay, and the content rules for new SKUs. You keep a platform (Salla, Zid, Shopify, or Woo). You do not keep the stacked hacks.

A fix means we stay inside the current theme: compress, defer, rename the CTA, turn off apps, correct mixed content, tighten titles. Same navigation, same staff habits, faster path.

Signs a theme fix is enough

Use this list as a go or no-go. If most rows are true, do not rebuild.

• The platform already matches how you sell (Salla or Zid for local payments, Shopify for a larger app graph, Woo if your team already runs WordPress well).
• Product data is clean: one URL per SKU, real photos, prices that match inventory.
• Checkout completes for you on a phone, even if it feels slow.
• RTL is mostly correct: price and quantity sit in a readable order.
• The red findings cluster on images, fonts, and extra scripts.

That cluster is the same work as Salla Zid Shopify speed base: compress photos, cut apps, one Arabic font. It is cheaper than a migration. It also shows up in why speed loses sales.

A fix still needs an owner. If no one can publish theme changes, a rebuild will stall for the same reason. Budget people, not only pixels.

What we actually change in a fix

Typical sprint: WebP pipeline, width and height on the LCP image, app embeds delayed, one primary Buy in Arabic, contrast pass, SSL check. Then a rescan.

We will not promise a 100 score by installing five "optimizer" plugins on Woo. We will not hide a broken checkout behind a new homepage slider.

Signs you need a store from scratch

Rebuild when the theme is the product and the product is broken.

• The theme was built for LTR English. Arabic labels clip, chevrons point the wrong way, and the sticky bar covers the CTA.
• Checkout is a museum of apps: one for coupons, one for upsell, one for local pay, none of them agreeing on order of fields.
• Product pages are duplicated (copied sections, copied descriptions, filter URLs indexed).
• Shipping and tax rules live in chat messages instead of the cart.
• Every "small edit" needs a freelancer who is afraid to touch `theme.liquid` or the Salla theme JSON.

That is not a PageSpeed problem. That is architecture. A store from scratch unifies identity, catalog, and one RTL checkout path.

Platform choice is size, not fashion

We pick Salla, Zid, Shopify, or WooCommerce from operations, not from Twitter.

• Salla and Zid when the catalog and payments are KSA/GCC first and the team wants a store admin they already know.
• Shopify when you need a wider app graph, multi-market, or a team that already lives there.
• WooCommerce when content and commerce must sit on one WordPress stack you can host and staff.

A rebuild on the same platform is normal. You keep the domain, the gateways, and the staff login. You throw away the theme debt. Migrating platforms at the same time as a UX rewrite doubles risk. Only do both when the current platform truly cannot pay or ship.

Cost, time, and what you keep

A theme fix is days to a few weeks if apps cooperate. A store from scratch is a project: sitemap of templates, product sample, checkout test on a real phone, redirects if URLs change, analytics re-fired.

Keep from the old store:

• SKUs, prices, and photos that are actually good (then compress them).
• Payment and shipping contracts.
• Reviews if you can export them cleanly.
• Domain and HTTPS once they are valid.

Throw away:

• Unused apps and their leftover scripts.
• Copied product text that Google already treats as duplicates.
• English-only section names that staff never understood.
• Popups that fire on first land from ads.

Read the scan report to mark keep versus drop. Even if you will rebuild, the report tells us which images and which URLs still convert.

SEO when you rebuild

If URLs change, map 301s. If they stay, you still rewrite titles so the new templates do not ship empty meta. Follow a basic crawl discipline like Google's SEO starter guide: index products, not every filter.

LCP still matters on the new theme. A pretty rebuild that ships a 4000px slider fails LCP the same way the old theme did. Speed is part of the build brief, not a phase after launch.

A decision rule you can use this month

1. Scan the advertised URL.
2. If high findings are images and apps, fix the theme.
3. If high findings are RTL, checkout maze, and cloned pages, build a store from scratch.
4. Do not migrate platforms unless pay or ops force it.
5. Do not buy a template pack. You need a path, not a skin.

If you are between the two, we still start with the scan. Patch the LCP image this week so ads waste less, then schedule the build. That hybrid is honest. Pretending a slider theme is a brand system is not.

Talk to x8x with the report in hand. We will say fix or build in plain language.

A fix week versus a build month

If you chose a theme fix, week one is not a rebrand. It is a narrow path: the ad URL, the LCP image, two extra apps, one Arabic button, the HTTPS lock. End the week with another scan. If the score moved and checkout completes on a real phone, spend week two on the gallery and fonts. If every change breaks something else, stop. That is the sign you need a store from scratch, not a third sprint.

If you chose a build, month one is not a slider. It is templates: home, collection, product, cart, pay, policy. A sample of ten real products, not lorem. A pay test with a card or a local gate on 4G. Image rules written before the photographer delivers 4000px files.

Content writes product titles in the same week the theme is built. Do not wait until after launch. A launch with empty titles sends you back to a red SEO row on the scan.

x8x prices those as two tracks. We do not mix a photo fix with a full catalog rewrite on the same invoice unless you ask. The scan is what splits the invoice.

Do not migrate mid-build

Switching platforms in a build week doubles tests: gates, shipping, staff, redirects. Keep the platform unless current pay cannot take your volume. Salla, Zid, Shopify, and WooCommerce all hold a clean store. The debt is in the theme, not in the company name in the footer.