Online retail stopped being a separate channel some time ago. It is now simply where a large share of shopping happens, and the interesting questions have moved on. Not whether people buy online, but how they find products, who or what places the order, and which parts of the experience actually change buying behaviour.
The numbers give a useful baseline. US retail e-commerce sales reached $340.2 billion in the second quarter of 2026, up 12.2 percent year over year and equal to 17.1 percent of all retail sales, according to the US Census Bureau. E-commerce is growing roughly twice as fast as retail overall, but it still accounts for well under a fifth of spending. That gap is where most of the opportunity sits.
Several forces are reshaping that landscape at once: mobile-first buying journeys, AI systems that both personalise and now transact, immersive product visualisation, sustainability rules that carry legal weight, and logistics networks that keep compressing delivery windows. This guide walks through each, with the evidence behind it.
Key Insights
- E-commerce is 17.1 percent of US retail sales and growing at roughly double the pace of retail overall (US Census Bureau, Q2 2026).
- AI has moved past recommendations: assistants can now complete purchases through open checkout protocols.
- Sustainability claims are becoming regulated data, not marketing copy, as the EU Digital Product Passport phases in.
- Drone delivery has left the pilot stage and is scaling to hundreds of US cities.
- Privacy compliance now means twenty state regimes in the US alone, not a single national rule.
Where E-commerce Actually Stands
It helps to separate the durable trend from the noise. Online retail’s share of US spending has climbed steadily for two decades, and Census data shows that climb continuing through 2026 without dramatic jumps. Growth is compounding, not exploding.
That matters for planning. A business that assumes online demand will double overnight tends to over-invest in capacity and under-invest in retention. The steadier reading suggests something else: the winners are gaining share from other online sellers, not just from physical stores. Competition inside the channel is the real pressure, which puts customer retention and margin discipline ahead of raw acquisition.
Category matters too. Books, electronics and apparel shifted online early. Groceries, furniture and anything requiring fit, inspection or installation moved far more slowly, which is exactly where blended physical and digital retail formats are being tested hardest.
The Rise of Mobile Commerce
Mobile is no longer a separate segment to optimise for. For most retailers it is the default context in which a customer first meets the brand, compares options and, increasingly, checks out.
How Mobile Changed the Buying Journey
The important shift is not the device but the fragmentation it created. A single purchase now regularly spans a social feed, a search result, a messaging app and a retailer’s own site or app. Attribution gets harder; so does maintaining a coherent story about the product across all those surfaces.
Practical consequences follow. Page weight and load speed matter more on mobile networks than on desktop. Checkout friction that a desktop buyer tolerates causes abandonment on a phone. Saved payment credentials and wallet support are no longer optional conveniences, and payment technology choices now directly shape conversion rates.
What Mobile-First Means for Store Design
Designing mobile-first means deciding what to cut. The product page that works on a phone shows one clear image, one price, one call to action and the two or three facts that resolve the buyer’s main doubt. Everything else moves below the fold or into an expandable section.
It also means treating the app question honestly. An app earns its place when repeat purchase frequency is high enough to justify the download, which is why grocery, pharmacy and subscription businesses invest in them while occasional-purchase categories usually should not.
Personalization and AI in E-commerce
Personalisation has matured from a differentiator into a baseline expectation. What has changed is the quality of the underlying models and, more importantly, the data they are allowed to use.
How AI Supports Tailored Shopping Experiences
AI-driven marketing has reshaped how brands reach customers, and recommendation systems sit at the centre of it. Modern systems combine browsing behaviour, purchase history and contextual signals to rank what a shopper sees. Done well, this reduces the work of finding the right product, which is the actual benefit. Done badly, it produces the familiar failure of being shown a product you already bought.
The more useful applications tend to be less visible: better search relevance, smarter size and fit guidance, restock timing, and pricing that responds to demand without eroding trust. Retailers building this out should look at how personalisation works in an e-commerce context before buying a platform, because the constraint is usually data quality rather than model sophistication.
Machine Learning and Customer Engagement
Beyond the storefront, machine learning shapes email timing, ad spend allocation and service routing. AI chatbots now handle a meaningful share of pre-purchase questions, and the good implementations hand off to a human at the point where the model starts guessing.
There is a governance dimension here that is easy to skip. Personalisation depends on data the customer has some claim over, and the regulatory environment (covered below) has tightened around exactly that. Building on first-party data and information customers volunteer directly is now a resilience decision as much as a compliance one.
Agentic Commerce: When the AI Places the Order
The most significant recent change is that AI assistants have started completing transactions rather than just suggesting them.
How Instant Checkout and the Agentic Commerce Protocol Work
In September 2025, OpenAI launched Instant Checkout, which lets a user buy a product without leaving the chat. It runs on the Agentic Commerce Protocol, an open standard OpenAI developed with Stripe and merchant partners. Etsy went live first for US users, with Shopify merchants including Glossier, SKIMS, Spanx and Vuori following.
The mechanics matter for retailers. The assistant passes order details to the merchant’s own backend; the merchant accepts or declines, then handles payment, fulfilment and support with existing systems. Merchants keep the customer relationship and pay a fee per completed purchase. Google has proposed a parallel standard, and card networks have published their own agentic payment frameworks, so the protocol layer is still being contested.
What Retailers Should Prepare
Two things are worth doing now, regardless of which standard wins. First, make product data machine-readable and accurate: structured attributes, real stock levels, unambiguous variant naming. An agent cannot buy what it cannot parse. Second, decide where the brand experience lives if the purchase itself moves off your site, which is a harder strategic question than the integration work.
This is also where conversational commerce finally delivered on a promise voice shopping never quite kept. Voice assistants proved good at reordering known items and poor at discovery, largely because comparing options without a screen is difficult. Text-based assistants sidestep that limitation, though optimising for voice and natural-language search remains relevant for how products get found in the first place.
Virtual and Augmented Reality Technology
AR in retail has settled into a narrower and more useful role than the early hype suggested.
Enhancing Consumer Experience with AR and VR
The technology answers one specific question well: will this fit, and how will it look in my space. Furniture placement, paint colour, eyewear, cosmetics and footwear are the categories where augmented reality in e-commerce earns its keep, because uncertainty about appearance is the main thing stopping the purchase.
Smartphone AR carried this, not headsets. The camera and depth sensing in an ordinary phone are enough for placement and try-on, which removed the hardware barrier that held earlier attempts back.
Where AR Actually Pays Off
The measurable benefit usually shows up in returns rather than conversion. Fewer wrong-size or wrong-colour orders means lower reverse logistics cost, and in apparel that cost is often the difference between a profitable and unprofitable category.
The honest counterpoint: AR is expensive to produce well. Each product needs an accurate 3D asset, and a catalogue of thousands makes that a serious commitment. Most retailers get better returns applying it to a high-value subset than to everything.
Sustainable Shopping Practices
Sustainability in commerce has split into two distinct things: what customers say they want, and what regulators are about to require. Both matter, but they operate differently.
Consumer Demand for Eco-Friendly Products
PwC’s Voice of the Consumer Survey, covering more than 20,000 people across 31 countries and territories, found consumers willing to pay an average of 9.7 percent more for sustainably produced or sourced goods, with 80 percent saying they would pay some premium. PwC itself notes the caveat: stated willingness and actual spending diverge under cost-of-living pressure.
Treat that 9.7 percent as a ceiling on the premium a green claim can support, not a floor. The businesses doing well here tend to compete on durability and repairability rather than on labelling, which is why resale and refurbished channels and circular business models have grown faster than premium eco-lines.
Transparency Becomes a Legal Requirement
The EU’s Ecodesign for Sustainable Products Regulation introduces the Digital Product Passport, a structured record of a product’s materials, origin and circularity data. Per the European Commission, the DPP registry framework becomes operational in July 2026, with the passport mandatory for certain battery categories from February 2027 and textiles, aluminium, tyres and construction products following between 2027 and 2029.
For anyone selling into the EU, this converts supply chain transparency from a marketing choice into a data-management obligation. Sourcing information has to be collected, verified and made retrievable at product level. That is a systems problem, and it is why supply chain sustainability now sits closer to operations than to communications.
Social Commerce Growth
Social platforms have become a genuine discovery and transaction layer rather than just a traffic source.
How Social Platforms Are Evolving
The pattern is consistent across markets: video-led discovery, in-app checkout, and creators acting as the merchandising layer. What differs is depth. In parts of Asia, live shopping is a mature primary channel; in the US and Europe, in-app purchasing has grown quickly but still complements rather than replaces retailer-owned storefronts. eMarketer projects US social commerce sales will pass $100 billion in 2026.
The structural trade is the same one agentic commerce raises. You gain reach and a shorter path to purchase; you give up ownership of the customer relationship and much of the data behind it. Our breakdown of how social commerce is developing goes deeper into that trade-off.
Creators as the Merchandising Layer
Influencer marketing has professionalised into something closer to a distribution channel with measurable performance. Affiliate-style attribution, dedicated creator storefronts and long-term partnerships have replaced one-off sponsored posts, mainly because one-off posts were hard to attribute and easy to fake.
The persistent difficulty is measurement across platforms, which is why brands running this seriously treat it as part of omnichannel marketing rather than as a standalone budget line.
Advancements in Supply Chain and Logistics
Delivery speed and reliability have become part of the product. Customers judge the purchase by when it arrives and how easy it was to send back.
Why Logistics Decides Margin
Last-mile delivery is the most expensive segment of the journey, and it is the part customers notice. The operational responses are unglamorous: better demand forecasting, inventory positioned closer to buyers, automated warehouse picking and clearer promise dates at checkout.
Resilience has moved up the agenda too, after several years in which port congestion, shipping-lane disruption and tariff changes all hit within the same planning cycle. Building supply chain resilience now means holding deliberate redundancy rather than optimising purely for cost, and cross-border sellers feel this most sharply.
Drone Delivery Moves From Pilot to Network
Drone delivery spent a decade as a demonstration. It is now scaling. Amazon has announced that Prime Air will reach nearly 500 US cities and towns by the end of 2026, up from 11 locations across 10 metro areas, with launches planned in Chicago, Cleveland, Atlanta, Syracuse and Boise. Orders can arrive in as little as 30 minutes, and Amazon says over 60 percent of its frequently purchased items are eligible by weight and size.
The constraint is physics, not ambition: small, light, high-frequency items in low-density areas. Groceries, pharmacy items and small household goods fit; furniture does not. Ground-based autonomous delivery is developing along a parallel track for denser urban routes.
Data Privacy and Security
Trust is a commercial asset in e-commerce, and the rules governing it have multiplied.
The Impact of Data Breaches on Consumer Trust
A breach damages a retailer twice: direct remediation and regulatory cost, then a slower erosion of willingness to share data. That second effect is the more expensive one, because personalisation, retention and lifetime-value modelling all depend on customers being comfortable handing over information.
The practical defences have not changed much. Collect less, retain it for a shorter period, encrypt properly, segment access, and be specific about what is stored and why. Vague privacy language reads as evasion.
The Compliance Map Keeps Growing
There is still no comprehensive US federal privacy law. Instead, twenty states now have comprehensive consumer privacy statutes on the books, with Indiana, Kentucky and Rhode Island taking effect on 1 January 2026. Each has its own thresholds and consumer rights, and further states have passed laws with later effective dates.
For an online retailer, the workable approach is to build to the strictest regime you are subject to and apply it broadly, rather than maintaining separate data flows per jurisdiction. Our overview of current data privacy developments covers how businesses are structuring that.
What to Watch Next
Three things are worth tracking closely over the next year.
First, whether agentic checkout consolidates around one protocol or fragments across competing standards. That determines how much integration work retailers face and how much leverage the assistant platforms hold.
Second, how marketplace dynamics shift as AI intermediaries insert themselves between shopper and seller. Marketplaces already compress margins; another layer of intermediation compresses them further.
Third, whether the sustainability regulation now phasing in changes product design or simply adds reporting overhead. The answer will differ by category, and businesses selling direct to consumers have more control over that outcome than those working through long distribution chains.
Conclusion
E-commerce in 2026 is a maturing channel, not a frontier. Growth is steady rather than explosive, which shifts the competitive question from capturing new online demand to taking share from other online sellers. That favours operational quality: accurate data, dependable delivery, honest product information and a checkout that works on a phone.
The genuinely new development is AI moving from recommendation into transaction. It is early, the standards are unsettled, and the strategic implications for brand and customer ownership are unresolved. Retailers who get their product data in order now will have options later. Those who wait will find the decisions made for them.
Found this useful?
Make SmartKeys a preferred source on Google, and our articles will surface more often in your Top Stories, AI Overviews, and AI Mode.
Add as Preferred Source
