In these interviews a few dozen companies describe a completed personalization move or one already underway, and most of the companies running an engine describe none — an observation from interviews, not a measured switching rate. Two things set this category apart from the CRM, CMS and social tool moves we have tracked. About a third of departures have no new standalone personalization vendor identified at the time of the interview, and the largest of those routes is the company's own build, usually at a large retailer that tested it against the vendor first. And a move here more often starts with the tool itself: capability is the complaint cited most often, ahead of cost and the effort to run it. Salesforce's bundled personalization is left more often than any single product. The companion reads are what personalization buyers wish they'd known, where the ratings live, and top complaints about personalization engines.
The pressure that builds: why companies switch personalization engines
The complaint buyers cite most often about the engine they leave is capability — in about half of the moves that name one — followed by cost and the effort it takes to run. A single move can carry several reasons, so the shares overlap. These are the pressures; the events follow.
It could not do the personalization the team wanted
The complaint cited most often in completed moves. A large specialty retailer moved off Adobe Target because “within Adobe Target, we were using a lot of third-party tools to try to do that personalization because we just weren't having enough offerings.” An apparel retailer took recommendations off Salesforce and onto Dynamic Yield: “The early data suggests we've seen some improvements versus Salesforce.” Another buyer had started “using Magento's Genie platform and just weren't impressed by what they had currently.” Sometimes the better option is the company's own: a large retailer says “we do see better performance with our in-house models than what we were seeing from this vendor.”
Named in this context: Adobe Target · Salesforce · Magento · Monetate · Dynamic Yield
The cost, set against what it delivered
Second to capability, and usually paired with it or with the effort of running the tool. One company that dropped its engine without a replacement found “It was hard to justify the cost and the time and effort it took to manage the system itself.” On the arriving side, a buyer sums up the choice: “So, the two things are cost and personalization.” The price-led moves themselves, and what the tiers cost, are on what buyers actually pay for personalization software.
Named in this context: Dynamic Yield · Salesforce
Too much work to run
An engine that needs more people than the team has. Of one testing-heavy tool, a buyer says “So it's self serve, any experiment, you're running it yourself, so it's time heavy.” Another company that left Dynamic Yield put it plainly: “this tool's just too expensive, it's too hard to use.” A third decommissioned its engine “so that we could invest in a new tool which is easier to use and requires less involvement from tech companies.” Dynamic Yield is both left and joined in these interviews: most of the companies leaving it are smaller and cite cost or effort, and most of those arriving are larger — a question of fit that the lessons page also draws.
Named in this context: Dynamic Yield · Qubit · Visually.io · Trbo · Yieldify
No proven lift, or no one using it
A few moves follow a test the engine lost or a tool that sat unused. A media company tested Salesforce's personalization in its campaigns and found “It didn't seem too enhanced in terms of personalization, and the results that we were getting didn't seem to be satisfactory.” Elsewhere, engines are simply retired: “these were some legacy platforms that were in there and just haven't fully gotten rid of.” Proving the lift is the hardest part of running these tools at all, which the lessons page covers.
Named in this context: Salesforce Interaction Studio · Optimizely
The event that fires the decision
In about half of the moves with a stated reason, buyers describe a broader organizational or vendor development alongside the product considerations: a vendor changing its offering, consolidation onto an existing suite, a replatform or a leadership decision to build internally — about the same share as in marketing automation, and fewer than in CRM, CMS or social tool moves. These are triggers or surrounding circumstances, not necessarily events outside the company's control, and they can overlap with capability or cost complaints. Vendor changes and suite consolidation account for most of them: without those two, about a quarter of the moves with a stated reason involve the remaining developments.
The vendor changed under the buyer
About one departure in six follows the vendor being acquired, a product being retired, or a vendor leaving the category. One vendor left website personalization altogether, and a B2B technology company describes what happened to its contract: “They decided that the future of their business wasn't in the website personalization piece, so they have sunset that piece from their product completely and basically handed over that half of the contract to VWO.” A retailer on a product that had changed hands describes where it stood: “We can only use what we build, we were not able to build new stuff.” Some of these buyers were satisfied with what they had to leave.
Named in this context: VWO · Monetate · Adobe · Constructor
Onto a suite, or a licence already paid for
Consolidation onto a platform the company already uses. One retailer “partnered with Salesforce a few years ago in an effort to move things in-house: our CDP platform, our email trigger program, our onsite personalization program.” Another set out to “unlock all the opportunities we have already bought from Adobe, which is Journey Optimizer, which I mentioned earlier and the Adobe Target.” Most of Adobe's and Salesforce's arrivals in these interviews come this way rather than as competitive wins.
Named in this context: Salesforce · Adobe Target · Salesforce Einstein
A replatform carried it
A new commerce platform or CMS brings its own recommendations, and the standalone engine goes with the old site. One brand weighing its options said “It's still an open decision whether we are going to transition that to Shopify or bring a vendor on board to help us with product recommendations” — and a year later ran Shopify's own.
Named in this context: Shopify · Salesforce Commerce Cloud · Adobe Commerce · Optimizely
A leader who wanted to own it
A few builds started with one person's conviction. At a DTC apparel brand, “Our head of engineering felt strongly that we needed to have homegrown to have the flexibility to do lots of different expressions of product recommendations on the site.” At a large retailer, “There was a leadership decision, which I don't disagree, to replace an off-shelf recommendation engine with our own internal ML based” — a story told below.
Named in this context: Nosto · in-house builds
Where do companies switch to?
Most observed departures go to another personalization vendor, and arrivals are spread across many of them, with no single destination dominating. Roughly a third have no new standalone vendor identified at the time of the interview: about one in six involve an internal build, about one in ten existing platform capabilities, and about one in ten have no replacement in place yet — a group that may still select a vendor. Salesforce's bundled personalization — Einstein, Interaction Studio, Marketing Cloud Personalization — is left more often than any single product, usually for capability, and its arrivals are mostly consolidations onto a suite the company already runs; one company that consolidated onto it is now moving back to the engine it left. Several products that were acquired or retired by their owners appear only as origins. These are directional findings from the interview set, not measures of vendor growth or churn. The leading engines are compared on Dynamic Yield vs Monetate and Optimizely vs Dynamic Yield.
Do companies build their own?
More often than in the other categories we track. About one departure in six ends in the company's own build — RichRelevance, Salesforce's recommendations and Nosto among the products replaced — against far fewer in CRM, CMS or social tool moves. Most of the builders are large retailers, and several tested the build against the vendor before cutting over: “We've been piloting our own in-house for the last three or four years because obviously we didn't want to just switch over.” An online retailer that moved from Salesforce's recommendations to its own heard from customers who wrote in to praise the new recommendations. The number of builds is small, so this is a distinctive pattern in these interviews rather than a market-wide rate, and it is not the usual route: most leavers still buy another vendor.
How long does a switch take?
Buyers provide few complete timelines from decision to cutover, so these interviews do not establish a typical duration for switching personalization engines. The examples instead show how different the work can be. The apparel retailer above had Dynamic Yield's recommendations live for about six weeks when it reported early improvements, although that does not establish the full migration timeline. A large retailer piloted its own recommendation engine for three or four years before replacing the vendor. Another buyer, weighing a renewal, described a product “that's taken us almost a year to implement.” Some transitions run two engines on the same site for a while, and a commerce replatform can set the cutover date. The time required depends on implementation, testing and the architecture involved rather than on a standard switching period.
Want a review of your own stack?
Get my review →The stories behind the switches
The counter-current: why most stay
Most companies running a personalization engine in these interviews describe no move at all, and a large share of those who looked are still evaluating rather than gone. Some keep their engine after a head-to-head — one retailer renewed its recommendations because “Certona drove incremental revenue, Fast Simon didn't.” Long tenure holds others: “as a company as a whole, we've been using this tool for so many years,” says one long-time Adobe Target customer, and another team's rule is “if it's not broken, we really try not to touch it.” Some re-evaluate and stay for now: “I think it sounds like for now we're planning to stick with Interaction Studio, but that's a conversation that I feel like is continually being reevaluated by leadership.” And some leave tools they like — for a consolidation, as one buyer moving off Adobe Target put it, “even though we are very satisfied with Adobe Target.”
What this means if you're weighing a switch
Buyers report that capability is the complaint most often behind a move, that about a third of documented departures have no new standalone vendor identified, and that the builds come mostly from large retailers that tested them first. Our recommendations, not practices buyers describe: before replacing an engine, list the personalization you actually run, since an engine that is under-used or a feature of a platform you already license may cover it; if you are considering a build, test it against the vendor on your own traffic and decide in advance what result you will accept, since one buyer here shipped a build that lost; if a consolidation onto a suite is driving the move, price in the people who will run it, which is the cost one buyer worried had been left out; and if your vendor has been acquired or is changing its product, start evaluating before the contract forces the date.
Common questions
Why do companies switch personalization engines?
In these interviews the complaint most often behind a move is capability — the engine could not do the personalization the team wanted — followed by cost and the effort it takes to run. In about half of the moves with a stated reason, buyers also describe a broader organizational or vendor development — most often the vendor being acquired or leaving the category, or a consolidation onto a suite the company already uses. These are triggers, not necessarily events outside the company's control, and the share is smaller than in the CRM, CMS and social tool moves we track.
Where do companies go when they leave a personalization engine?
Most move to another vendor, and arrivals are spread across many of them. About a third have no new standalone vendor identified at the time of the interview: they build their own, switch on a feature of a platform they already have, or have no replacement in place yet — and that last group may still choose one. These are observed decisions in the interview set, not churn rates or market share.
Do companies build their own personalization or recommendation engine?
Some do: about one departure in six in these interviews ends in the company's own build, more often than in CRM, CMS or social tools. Most of the builders are large retailers, and several ran their build against the vendor before switching.
Is the personalization bundled in Salesforce or Adobe enough?
Salesforce's bundled personalization tools are left more often than any single product in these interviews, usually because they could not do what the team wanted, and the moves onto them are mostly consolidations. Adobe's arrivals are mostly companies putting a licence they already hold to use. Whether a bundled tool is enough depends on how much personalization the team actually runs.
How long does it take to switch personalization engines?
These interviews do not establish a typical switching time: few buyers give a complete timeline from decision to cutover. The examples vary widely — one retailer had a new vendor's recommendations live about six weeks before reporting early results, another piloted its own engine for three or four years, and one buyer described an implementation that took almost a year. Implementation, testing and the architecture involved drive the time.
What happens when a personalization vendor is acquired or leaves the market?
Several moves in these interviews began that way. Buyers moved to the acquirer's other product, to a partner the vendor pointed them to, to their own build, or off the tool entirely — and some were satisfied with the product they had to leave.
A tech stack review, powered by the buyers who've been there
What matters most to you, a shortlist for your stack, and the next steps to take — built from hundreds of verified buyer interviews.
Get my stack reviewFree · no password · your answers are anonymized before they inform research like this.