A Shopper Browses Twice and Leaves. Which Platform Wins Them Back?

Every marketing platform says it personalises in real time. So here's a simple test. A shopper looks at the same product twice in two days, doesn't add it to the cart, and leaves. How fast can your team turn that into a live, on-brand message on the right channel, without moving data between tools?

That moment is called browse abandonment, and it's one of the clearest signals of intent a website gets. Most platforms can respond to it. What separates them is how many products, licences and people it takes to get there.

We put six platforms through the same test, using each vendor's own documentation (checked 11 November 2026). We make one of them, so we've given ours the most cautious write-up.

The test, step by step

To win the shopper back, five things have to happen:

1. Catch it. Record both product views against the same person.
2. Know who it is. An email or SMS needs a known contact who has agreed to hear from you.
3. Decide. Is two views enough to act on, and is this person already in another campaign?
4. Write it. An on-brand message about that product, not a generic "come back" email.
5. Send it and stop it. The right channel at the right time, and nothing more if they buy.

Each platform below is judged on those five steps, plus one question: how many tools does the data pass through on the way?

HubSpot: strong, if the shopper is already a contact

HubSpot workflows can enrol a contact when they view a specific page, and Marketing Hub Professional and Enterprise can track custom events such as a product view (HubSpot).

The catch is step 2. HubSpot's own documentation says a page view won't trigger the workflow if the visitor hasn't yet become a contact. Custom events only enrol existing records too. If a visitor triggers the event and becomes a contact a minute later, they still miss the workflow.

Passes if the shopper is already in your HubSpot CRM and your product data lives there too.

Struggles when most of your browsers are anonymous, or your store data sits in another system.

Salesforce Marketing Cloud: fast, on the right version

Salesforce's newer architecture, Marketing Cloud Next, is built on its customer data platform, Data 360. According to Salesforce, a page view is ingested as a streaming event and updates the customer's profile. A Data 360 Triggered Flow can then send an email or SMS within minutes, and an Agentforce agent can choose the content, offer or channel.

That's a strong answer to the test. But Salesforce's own page also says the older Journey Builder, built on Marketing Cloud Engagement, relies on list-based entry and scheduled audience builds. So the real answer depends on which Marketing Cloud you run, and whether Data 360 is connected to your live website data.

Passes if you're on Marketing Cloud Next with Data 360 receiving live web events. Struggles when you're on the older setup and "real time" means waiting for the next audience build.

Adobe Real-Time CDP: best in the session, but it doesn't send the email

Adobe is strongest at step 1. Real-Time CDP streams behaviour through Adobe's Edge Network and can change what a visitor sees on the same page or the next one, in the same session. For many shoppers that's the best answer of all: the website responds before they leave.

But Adobe is clear that Real-Time CDP activates audiences, while a separate application, Adobe Journey Optimizer, runs the messages and decides what happens next for each person (Adobe). The two share one customer profile, so no data moves between them. Producing creative at scale is the job of a third product, GenStudio.

Passes if you already run the Adobe stack. Struggles when you bought the CDP and expected it to send the follow-up email.

Klaviyo: built for exactly this, for known shoppers

Klaviyo has the most direct answer. It offers a pre-built browse abandonment flow, triggered by its Viewed Product metric, that sends email or SMS (Klaviyo). Klaviyo Academy even suggests a filter so the flow only sends after someone has viewed a set number of products in a set time, which is close to our "twice" test.

The catch, again, is step 2. Klaviyo only tracks browsing by identifiable visitors: people who've clicked through a Klaviyo email or SMS, submitted a Klaviyo form, or logged in where custom tracking is installed (Klaviyo). If anonymous visitor tracking is turned on, it can store a visitor's activity and add it to their profile once they're identified.

Passes if you sell online and most of your browsers are already on your list. Struggles when you're a B2B team, or most of your traffic is new.

Jasper: writes the message, can't see the shopper

Jasper isn't built for this test, and that isn't a criticism. It's a content platform with strong brand controls (Brand Voice, Visual Guidelines and Style Guide) and agents for marketing workflows (Jasper). It could write a very good on-brand reminder for step 4. (Our guide to keeping AI content on brand covers what to check in any content tool.)

It isn't a CRM or a customer data platform, though. Catching the browse, knowing the shopper and sending the message all happen in other tools, and the content moves between them.

Passes if step 4 is your bottleneck and the rest of your stack already works. Struggles when you need one platform to run the whole test.

GalaxiQ: one platform for all five steps, with a shorter track record

Here's how GalaxiQ is designed to handle the test, from our EdgeIQ product page. EdgeIQ captures every visit, click, cart and purchase as it happens. Atlas, its AI segmentation agent, turns those events into audiences such as Returning Shoppers. You describe the journey in plain English, and EdgeIQ builds it and picks the channel for each step: email, on-site popup, SMS or video.

The content for each segment comes from Ideanise, written in the brand voice set by BrandForge by GalaxiQ, inside the same platform. The data doesn't move between tools. (More on how the products fit together in what is GalaxiQ?)

Now the cautious part:

• GalaxiQ was founded in 2025, so it has a shorter track record than every other platform here.
• The email and SMS steps still need a known shopper who has agreed to hear from you. That's the same rule that catches HubSpot and Klaviyo.
• None of this should be taken on our word. Ask us to run this exact test on your own website before you sign.

The test at a glance

How to run browse abandonment well, whatever you use

These come from Klaviyo's own guidance and apply to every platform:

• Wait before you send. Klaviyo recommends a delay of at least an hour before the first message.
• Keep it lighter than a cart reminder. Viewing a product shows less intent than adding it to the cart, so the tone should be softer.
• Stop when they act. Exclude anyone who has started checkout or placed an order since entering the flow.
• Don't repeat yourself. Klaviyo's suggested filter keeps people out of the flow if they've been in it in the last 30 days.
• Act on real interest. Sending only after several product views improves engagement.
• Respect consent. In the EU, EEA, UK and Switzerland, some stores won't track visitors who haven't consented, so those shoppers never enter the flow.

The biggest limit on every platform is the same: anonymous visitors. Forms, logins and earlier email clicks are how a browser becomes a known shopper, so growing that known audience is the work that makes browse abandonment pay off. Our guide to AI customer segmentation covers which signals are worth acting on.

The question to ask in every demo

Don't ask a vendor whether they do real-time personalisation. They'll all say yes.

Ask them to run this test live. One test shopper, two product views, then leave. Count the tools the data touches, the people involved and the minutes until the message lands. If the answer involves an export, a ticket to the data team or "next sprint", the shopper has already bought somewhere else.