E-Commerce Personalization in 2026: Bloomreach, Algolia, Constructor, Coveo, and the Post-Cookie Reality
The search and personalization vendor stack sorted honestly — what Constructor does that Algolia doesn't, where Salesforce Personalization (formerly Interaction Studio) actually fits, and what Shopify Sidekick changes.
E-commerce personalization is the AI category where the vendor pitch and the operational reality diverge most sharply. The pitch — “one-to-one personalization driving double-digit revenue lift” — has been the dominant marketing message since roughly 2015. The reality in 2026 is more complicated, more interesting, and shaped by three forces the early personalization vendors largely missed: the slow death of third-party cookies, the rise of large-language-model-flavoured search, and the platform consolidation that put commerce engines, search engines, and personalization engines into the same buying conversation.
This is the honest sort of where the personalization stack sits — Bloomreach, Algolia, Constructor, Coveo, the Salesforce repositioning, and what Shopify Sidekick changes for the mid-market.
What personalization actually means in 2026#
Personalization in e-commerce splits into three operational layers, and most platform sales pitches blur them.
The first is product discovery — how search results, category page sort order, and the on-site search ranking adapt to the visitor’s behaviour. This is where most measurable revenue lift lives, and where the rebuilt vendors (Constructor, Algolia’s recent generation, Bloomreach’s Discovery product) compete hardest.
The second is recommendations — the “you may also like,” “frequently bought together,” and email carousels that surface incremental items. The technical baseline is well-understood and increasingly commoditised; the differentiation now is in the integration with merchandising rules rather than the underlying model.
The third is journey orchestration — the cross-channel pattern of when to send what to whom (email, push, on-site, paid retargeting). This is where Salesforce Personalization (formerly Interaction Studio) and Bloomreach’s broader engagement platform actually sell.
A retailer evaluating “personalization” without separating these layers ends up buying the wrong tool for the actual revenue problem.
Bloomreach#
Bloomreach (Czech-origin, US-headquartered, valued at 2.2 billion USD in its 2022 round, profitable as of late 2024) has the broadest product breadth in the category. Discovery (search and category management), Engagement (email, SMS, push, on-site orchestration), and Content (the headless CMS roots of the company) compose a platform that competes against three different vendors depending on the deal. The 2024-2025 push was the Loomi AI layer — generative AI for merchandiser productivity and conversational search interfaces.
Where Bloomreach wins: retailers that want one vendor for search, recommendations, and cross-channel orchestration, and have the data-engineering capacity to feed the platform well. Where it loses: deals where the customer already standardised on Salesforce Marketing Cloud or Adobe and wants a pure-play search-and-discovery tool.
Algolia#
Algolia (Paris-origin, US-headquartered, valued at 2.25 billion USD in its 2021 round) has rebuilt aggressively over the 2023-2025 window from a developer-loved search API into a full discovery and personalization platform. The pivot — partly forced by Constructor’s competitive pressure, partly enabled by the LLM-augmented search wave — added neural search, AI re-ranking, and a merchandiser interface that older Algolia customers had been quietly asking for.
Algolia’s strength is the developer ecosystem and the global infrastructure footprint. The weakness historically was that it was not a merchandiser tool — the merchandising team needed the engineering team to ship rule changes — and the 2024-2025 product investment substantially closed that gap.
Customers include Lacoste, Decathlon, Stripe (for their docs but also internal use), and a long tail of mid-market e-commerce. The 2026 question is whether Algolia successfully completes the platform transition or remains anchored as a search API in larger accounts.
Constructor.io#
Constructor (San Francisco-based, founded 2015) is the most aggressive recent disruptor in the category. The pitch is “AI-first search and discovery measured on revenue, not click-through” — a meaningful operational distinction because click-through optimisation can produce revenue-neutral or revenue-negative outcomes that the older personalization platforms quietly ignored.
The Constructor customer list includes Sephora, Petco, Backcountry, and a deep bench of mid-to-large specialty retailers. The technical credibility is high — the underlying reinforcement-learning models genuinely outperform the older relevance-tuned approaches on revenue-weighted metrics — and the merchandiser interface was designed around real e-commerce workflows.
Constructor’s weakness is the same as the strength: it is focused on search and discovery, not the broader engagement story. Retailers that want a single platform for everything will end up at Bloomreach or in the Salesforce/Adobe stack; retailers where search rank is the binding revenue constraint should be evaluating Constructor seriously.
Coveo#
Coveo (Montreal-based, public since 2021 on the TSX) covers an interesting middle ground — search and personalization for both commerce and the workplace (the knowledge-base-search side of the business). The commerce product is credible, the customer base is solid (Lenovo, Salesforce’s own help center, several large B2B distributors), and the 2024-2025 generative AI push integrated meaningfully with Coveo’s existing relevance pipeline.
Where Coveo wins: B2B commerce, distributors with complex catalogue structures, and retailers where the search-relevance problem is the binding one. Where it loses: pure B2C consumer commerce against Bloomreach or Constructor.
Salesforce Personalization (formerly Interaction Studio)#
Salesforce rebranded Interaction Studio as Personalization in 2023 as part of the broader Data Cloud and Einstein 1 platform consolidation. The product is real and the customer base is large among Salesforce Marketing Cloud customers, but the standalone competitive position outside the Salesforce stack has weakened over 2024-2025. The honest read: if you are a Salesforce customer with Data Cloud and Marketing Cloud already deployed, the Personalization product is the path of least resistance and increasingly capable. If you are not already in the Salesforce stack, you will compete-bid Personalization against Bloomreach and end up with Bloomreach more often than not.
Shopify Sidekick and what it changes for the mid-market#
Shopify Sidekick — the AI assistant Shopify shipped in beta in 2023 and broadly in 2024 — is reshaping the mid-market personalization conversation. For Shopify Plus merchants, the embedded AI handles a meaningful slice of the merchandiser workload (product descriptions, image alt text, basic recommendation tuning, store-front content variations) that previously required a personalization vendor add-on.
The 2025 Sidekick expansion added natural-language merchandising commands and tighter integration with Shopify’s recommendation models. For Shopify merchants below roughly 50 million USD in annual revenue, Sidekick plus the native recommendation features now covers the common personalization use cases without a separate platform. Above that revenue level, the conversation shifts to Bloomreach, Algolia, or Constructor depending on the discovery-versus-engagement balance.
The post-cookie first-party data shift#
The Chrome third-party cookie deprecation — repeatedly delayed but operationally inevitable through 2025 and 2026 — combined with Apple’s ATT framework and the broader Safari and Firefox privacy posture, ended the era of cross-site behavioural targeting as the default personalization signal. The 2024-2026 response across personalization vendors has been the first-party data push — customer data platforms, identity resolution against authenticated users, and on-site behavioural signal as the primary personalization input.
This is mostly good news for personalization vendors that historically built on first-party data (Bloomreach, Constructor, Coveo) and a headwind for retargeting-heavy vendors. The retailers that handled the shift well had an investment in their customer data platform — Segment (now Twilio), mParticle, Tealium, or the cloud-native equivalents — that produced the unified identity layer the new personalization stack needs.
What we recommend in 2026#
For a retailer scoping a personalization investment:
- Pure-play search and discovery as the binding constraint — Constructor for B2C consumer, Coveo for B2B or complex catalogue, Algolia if developer-led.
- One vendor across discovery and engagement — Bloomreach is the default; the cross-channel orchestration story is genuinely real.
- Existing Salesforce Marketing Cloud customer — Salesforce Personalization is the path of least resistance.
- Shopify Plus mid-market — Sidekick plus native recommendation features handle most of the workload; add a vendor only above the revenue threshold where the math starts working.
- Post-cookie identity layer — fix this before evaluating personalization platforms; the personalization platform without a working first-party identity is buying nothing.
The biggest mistake we see in 2026 personalization deals is buying the platform before fixing the data layer. The vendor demo always looks good with clean data; the production deployment quietly fails when the catalogue, customer, and behavioural data aren’t reconciled.
Related reading#
Where pdpspectra fits#
We help retailers stand up the first-party data layer underneath whichever personalization platform they pick — identity resolution, catalogue normalisation, behavioural pipeline, and the integration glue between the e-commerce platform and the personalization vendor. Our data engineering practice handles the operational layer.
Personalization without a working data layer is theatre. If you are scoping a search and discovery rebuild, or fixing a post-cookie identity gap, tell us about the storefront.