E-Commerce Trends We Got Wrong: Honest Reflections on Predictions

The e-commerce industry runs on predictions. Every January brings a fresh wave of "trends to watch," and we have written our share — see our forward-looking pieces on 2027 through 2030. But a forecast is only worth as much as the person making it is willing to grade it afterward. So this is the uncomfortable companion piece: the trends we, and most of the industry, called wrong — and, more usefully for you, what each miss should change about how you spend money on your PrestaShop store today. The pattern matters more than the individual predictions, because the same pattern is being repeated right now with AI.

We will be concrete about PrestaShop throughout. A "trend" only matters to a shop owner when it turns into a module to install, a setting to flip, or a chunk of developer time to fund — so for each one we will tell you exactly what the back-office decision was, and whether spending on it paid off.

Voice commerce: still a timer-and-weather machine

Around 2018–2020 the consensus was that voice shopping through Alexa and Google Home would reshape retail. The story was tidy: tens of millions of smart speakers in homes, frictionless reordering by simply saying what you want.

What actually happened: voice remains a rounding error in online sales. People use smart speakers for music, timers and the weather. Shopping is stubbornly visual — buyers want to see the product, scan reviews, compare options. "Alexa, order running shoes" with no sizes, styles or prices on screen feels reckless, so people don't do it.

What this meant for PrestaShop owners. The merchants who got burned were the ones who paid for voice-search "optimization" or speculative skills/actions integrations. The ones who came out ahead spent the same effort on structured product data that happens to help every channel: complete Catalog → Attributes & Features, clean Catalog → Categories hierarchies, and proper schema/rich data. None of that was wasted, because it serves ordinary search too. So what? When a hyped channel asks you to build something that only serves that channel, that is the warning sign. Investments that pay off regardless of whether the trend lands are the only safe ones.

AR/VR shopping: great demos, narrow reality

Augmented reality was supposed to let customers drop a sofa into their living room or try sunglasses on their face. The technology genuinely works, and the demos are impressive.

Reality: very few shoppers actually open an AR view. The friction — launch the feature, point the camera, wait for it to render — is higher than just looking at good product photos. The real exception is high-consideration, high-anxiety purchases: expensive furniture, custom kitchens, made-to-measure goods, where visualization removes a genuine reason not to buy.

What this meant for PrestaShop owners. If you sell £15 phone cases, an AR viewer was money lit on fire. If you sell £2,000 wardrobes, it can be worth it — but only after the fundamentals are in place. For most stores the boring win beat the flashy one: multiple high-resolution images, a working zoom, and short product videos delivered without wrecking your page speed (lazy-load them; see the e-commerce costs nobody talks about for how "free" media features quietly cost you in performance and bandwidth). Lesson: a technology can be genuinely useful for a narrow niche without being the broad transformation it was sold as. Ask whether your catalog is the niche before you pay for it.

Blockchain and crypto payments: a solution hunting for a problem

In 2021–2022, accepting crypto was framed as table stakes for any "forward-thinking" store. NFT loyalty programs, blockchain supply-chain tracking, decentralized marketplaces — all of it was going to upend commerce.

Reality: crypto payments for ordinary retail stayed negligible. The vast majority of buyers reach for a card or PayPal. Volatility makes crypto impractical for everyday purchases, and the user experience — wallets, gas fees, confirmations — is far too heavy for mainstream shoppers.

What this meant for PrestaShop owners. Every payment method you bolt onto checkout has a real cost beyond its install: it is another module to keep updated and secure, and another option that can clutter the payment step and slow a decision. A crypto gateway you process two orders a year through is pure liability. The discipline here is to add payment methods your actual customers ask for — usually a card processor, PayPal, and the local favourite for your market (BLIK in Poland, Klarna in the DACH region, and so on) — and to keep the payment step short, because a crowded payment screen costs you more conversions than the missing exotic method ever would. So what? Enthusiasm inside a tech niche is not a signal about mainstream behaviour. Watch what regular consumers reach for, not what early adopters evangelize.

Social commerce: real growth, no replacement

This prediction was half right, which makes it the most instructive. Social commerce — buying directly inside Instagram, TikTok and the rest — has grown substantially. But the bolder claim, that social platforms would replace standalone stores, did not happen. Most social commerce still pushes traffic to an external store rather than completing the sale natively.

The reason is trust. Shoppers still prefer buying somewhere they can navigate, with a clear return policy and real customer service, over a fleeting post. Social platforms are excellent for discovery and mediocre for transactions.

What this meant for PrestaShop owners. The merchants who treated social as a discovery channel feeding their own store made the right call. In PrestaShop that means a clean product feed for shopping ads and catalogs, UTM-tagged campaign links so you can see in your stats which channel actually pays, and a fast landing experience for the visitor who taps through. The merchants who tried to move their whole business onto a platform handed away their margin, their customer relationship, and their data. Your own store is the one asset you control — which is also the through-line of why customer lifetime value should drive your marketing: a follower you don't own is worth far less than a customer you can email. Lesson: new channels supplement existing ones far more often than they replace them.

Chatbots as sales agents

Service chatbots were supposed to handle sales conversations, answer product questions and guide purchases. The early ones were so bad they became a punchline — looping menus that never reached your actual question.

Modern large-language-model chat has genuinely closed that quality gap; the conversations are real now. But adoption as a sales tool is still limited, because most shoppers prefer to browse and research on their own terms. Chat earns its keep in post-purchase support far more than in pre-purchase selling.

What this meant for PrestaShop owners. The chat investment that paid off was the unglamorous one: routing genuine "where is my order / how do I return this" questions to a fast answer. The investment that disappointed was expecting a bot to close sales a good product page should have closed already. That is also why we still answer real tickets ourselves rather than hiding behind a deflection bot — see why we answer every ticket within hours. Lesson: sometimes the technology catches up but the use-case prediction was off by one. AI chat turned out to be a support tool wearing a sales-tool costume.

The honest scorecard: what we actually got right

Grading yourself only on the misses is its own kind of dishonesty. Several predictions held up, and they share a tell — none of them depended on consumers adopting a brand-new behaviour. They simply asked existing behaviour to grow.

Read down the "Right" column and the pattern is unmistakable: the winners amplified behaviour people already had. The losers asked people to behave in a new way. That single distinction would have correctly sorted almost every trend of the last decade — and it is the lens to keep pointed at whatever is hyped next.

Mobile and page speed: the calls that paid off in PrestaShop terms

Two of those right calls deserve a concrete word, because acting on them was free and the laggards are still paying for it. Mobile-first is now simply how stores are built: if your theme, your responsive breakpoints, and above all your checkout aren't tuned for a thumb, you are losing the majority of your traffic at the worst possible moment. Page speed turned from a nice-to-have into a Core Web Vitals reality you can see and fix — combine, minify and cache your assets under Advanced Parameters → Performance, lazy-load imagery, and keep an eye on the third-party scripts (chat widgets, trackers, that abandoned AR viewer) that quietly add seconds. These were the boring fundamentals all along, and they out-earned every flashy trend on the wrong-column list.

How to evaluate the next trend before you spend on it

After getting enough calls wrong, we tightened this into a checklist we run before funding anything new. Four questions, in order — if a trend fails an early one, the later ones don't matter.

• Does it solve a problem a real customer actually has? Exciting-for-technologists is not the same as painful-for-shoppers. Voice, crypto and AR all failed here for most catalogs.
• Is the friction lower than today's option? A new technology has to be easier than what it replaces, not merely different. Crypto checkout was harder than a card; that was the whole story.
• Are regular consumers using it? Not early adopters, not the tech press — ordinary buyers. Check real adoption data, not trend articles (including ours).
• Can you test it cheaply and reverse it? The right response to a promising trend is a small, undoable experiment, never a bet-the-store overhaul. PrestaShop makes this easy: trial a module on a staging copy, measure for at least 30 days and a few hundred orders, and keep or remove it on the evidence.

That last point is the difference between a discipline and a gamble. Don't trust a gut feeling and don't trust three days of data — record your real numbers (conversion rate, average order value, mobile-versus-desktop) before you change anything, then judge the change against them. Tie it back to economics rather than excitement: a trend is only worth funding if it improves a number you can see in your real profit margins.

The trend we're getting wrong right now: AI

The honest reason to write a retrospective is to avoid repeating it, and AI is the live test. Some of it is genuinely useful today — drafting product descriptions, smarter search and recommendations, triaging support tickets. Some of it is the 2021 crypto pitch in a new outfit: "add AI to your store" with no problem named. The four questions above sort the two cleanly, and we drew that line in detail in AI in e-commerce: what works today vs what is just hype.

The throughline across every miss on this page is the same: technology adoption follows how people actually behave, not how we imagine they should. The merchants who stayed solvent through a decade of trend cycles were rarely the first to chase the new thing. They were the ones who kept pouring money into the unglamorous fundamentals — fast pages, a clean catalog, a checkout that gets out of the way, reliable delivery, real support — and ring-fenced a small, reversible budget for testing what might be next. That is not a thrilling strategy. It is the one that was still standing when the smart speakers went back to telling people the weather.