Lite Versions and AMP: A Verdict for Small Stores

Do not build a separate lite version of your store, and do not adopt AMP for a small store's pages in the current platform landscape. Both promise speed by subtraction, and both deliver a second surface to maintain: prices, stock, offers and content drifting between versions, visitors crossing from the lite page to a heavy checkout and losing the benefit at the exact moment it matters, and a stripped experience that often removes the proof and detail the purchase needed. The speed these formats chase is available on the main site through the boring programme of this bench: images under discipline, scripts on a budget, a theme chosen for weight. One site, made properly light, is faster to run in every sense that bills you.

Lite versions and AMP were rational responses to an era of heavy templates, slow devices and search distribution that rewarded the formats directly. The ground has shifted: mainstream themes are far lighter, browsers and devices are faster, the distribution advantages have faded, and the major platforms build responsive, reasonably lean pages by default, as their documentation shows. What has not shifted is the maintenance arithmetic, which always favoured the format's promoters over its operators. Judge any such proposal against today's conditions and your own field data from the field data guide, not against the problem as it stood when the format launched. Nostalgia is not a performance strategy.

Every fact on a store lives twice in a two-version world: the price, the delivery terms, the stock state, the offer dates, the returns wording. Each fact will, in time, be updated in one version and not the other, because the person updating is busy and the second version is nobody's favourite child. Visitors meet the drift as a lite page promising what the checkout denies, which is worse than slowness: it is the message-match failure the landing bench warns against, manufactured by architecture. Even where generation is automated from one source, templates diverge in features and fixes. A small business has no spare attention for a second surface; spending it there starves the surface that takes the money.

Lite formats achieve weight loss by removing things, and the things nearest the knife are rarely the ones the audit in the script budget would choose: reviews, detailed imagery, size and fit tools, rich product information, the very content that converts a considered purchase. What remains is fast and thin, and thin pages bounce visitors who needed substance back to the full site, where they were headed anyway at the cost of a confusing detour. Speed is a means to a completed purchase, not a score to be maximised against the purchase itself. Where your main site carries genuine bloat, remove the bloat; do not build a lifeboat and leave the ship heavy.

Three exceptions deserve honesty. A business serving regions where devices and networks are severely constrained may justify a genuinely lightweight template, designed as the main experience for that market rather than as a stripped copy. Extremely content-led pages, a help centre or a publication arm, can adopt aggressive minimalism profitably because their product is text. And platform-provided accelerators that are part of the platform itself, maintained by the platform rather than by you, should be judged as features, not as a second site. Each exception shares a trait: the light version is somebody's main job, maintained as a primary surface. If yours would be a side project, it does not meet the exception's terms.

Spend the lite budget on the main site's programme, in the pillar's order. Images first, with the routine from the images guide. Scripts second, on a budget with an owner. Theme third, if the floor itself is heavy, judged by the evidence method in theme speed. Mobile layout and payment alongside, since speed without usability is a fast exit. That programme reaches the same visitor on the same page they will buy from, with no drift surface and no second set of promises to keep. It also compounds: every future page inherits the discipline automatically, while a lite estate must be re-explained to every future change.

When a lite or AMP proposal arrives, ask three questions in writing. What field-data problem, on which template, does this solve that the standard programme cannot? Who maintains the second surface weekly, by name, and what work of theirs does it displace? What happens when a visitor crosses between versions mid-purchase? Proposals that survive those questions are rare and usually turn out to be the exceptions above wearing proposal clothes. The rest dissolve, and the budget returns to the unglamorous work that the pillar guide sequences. Verdicts like this one exist to save a small business from a category of project entirely, which is the cheapest saving a bench can offer.