Comparisons

MandM, a British online fashion retailer, captured 88% more email signups from their popups. They didn't rewrite the copy. They didn't redesign their site. They just changed when their message appeared.

How did they manage this? Simply, they made the switch from rule-based experience delivery, to an intent-based approach.

And the results across email capture and product recommendations tell a consistent story: rule-based experience delivery forces a single answer on a question that has many right answers.

Rigid, predefined rules can't answer those questions. Intent signals can. Improving the impact of your onsite experiences is all about sending the right message, at the right time to your customers. We'll show you how Ollie Wilson, Insights Activation Manager at MandM does this with Made With Intent.

Editor's note: This blog post is a write up based on our first Intent Live: How MandM maximise revenue per user with personalised experiences. The session was hosted by Ollie Wilson, Insights Activation Manager at MandM. He showed how Made With Intent helped deliver better, more appropriate experiences to his customers, getting 88% more email signups.

The problem with rules-based personalisation

Most eCommerce personalisation sits on top of a set of rules. A customer views a Product Landing Page (PLP), then two Product Display Pages (PDPs), then gets hit with an email capture popup.

Or they get served "last viewed" recommendations based on browsing history. Or a basket abandonment email fires after 10 minutes of leaving the site.

These rules work to a point. Delivering the same experiences to every visitor using predefined rules, based on what they've done before gives you a critical foundation, but it puts a ceiling on growth.

They treat the journey as a sequence rather than a state. And a customer's state when they trigger your rules can be completely different depending on who they are, why they're there, and what they're about to do.

MandM saw this clearly in their email capture data. Their rule-based popup was capturing emails, but they were seeing broken journeys. The pop-up was technically firing at the right moment in the sequence. It wasn't firing at the right moment for the person and their intent.

As Ollie Wilson, Insights Activation Manager at MandM, puts it:

"It's not necessarily specific things a customer does in the journey. It's more so the timing and intent really helped us leverage this in a more efficient way."

The argument we're making here is that it's key to make this distinction. Rules track what a customer has done. The moment for you to intervene has gone. Intent-based experience delivery is issued in real time, predicts what they're about to do, and allows you to take appropriate action.

Pop ups delivered at the right time

The email capture popup is one of the highest-value tools in eCommerce, but also one of the most frequently misused. Use it too early and you interrupt a customer who hasn't found a reason to stay yet. Fire it according to a fixed rule and you'll hit some customers at the peak of their interest, but most others at exactly the wrong moment.

Ollie and his team tested a different approach. Instead of triggering the popup after a visitor hit a fixed sequence of pages, they introduced intent signals to identify when a customer was building meaningful engagement.

When those signals crossed a threshold, the popup fired. Exactly at the moment the customer was most receptive.

When talking about this, Ollie said: "We were hitting them at the right time because we knew they were building intent. They were right at the peak of their journey. Whereas before we were very much relying on this rule-based system which potentially wasn't the right time."

The results across three metrics tell the story. And these figures are lifted directly from the numbers Ollie shared during Intent Live:

- 55% increase in email sign-up rate, the rate at which people served the popup chose to subscribe

- 88% uplift in total emails captured, the volume consequence of that improved rate

- 15% resubscription rate among previously unsubscribed customers

That last number is particularly significant. Lapsed customers don't re-subscribe because of a well-timed popup by accident. They re-subscribe because they were caught at a moment of genuine brand or product affinity, at the right time. This simply isn't possible with rules-based personalisation.

What MandM actually changed

What's actually surprising is how little MandM had to do to arrive at these results.

They were already using Bloomreach for their on-site experiences, including the consent popup. The popup itself, design, copy, offer, stayed exactly the same.

On changes, Ollie says: "To be honest, it was so easy. We already had our consent popup in Bloomreach as a web layer. It was just a trigger we had to change. With Made with Intent's integration being so easy, we could just feed all of the data into Bloomreach and then use that as the trigger for the popup rather than those stringent rule bases."

Agentic campaigns: testing 157 segment combinations at once

The email capture result came from MandM's first phase of work with intent signals. Their second phase went further, introducing a different kind of challenge.

MandM runs an active personalisation testing programme. At any given point, they're running recommendation strategy tests: last viewed versus category affinity versus the Bloomreach Loomi engine versus most popular, and so on.

Each test runs for roughly two weeks, produces results for a specific segment or device type, and then the cycle starts again.

The problem isn't that the testing doesn't work. It's that it's slow. Each test answers one question, for one segment, in one context. And by the time you've worked through a few cycles, the results of the first test may not apply to the next season, the next acquisition cohort, or mobile versus desktop. And not to mention how resource intensive this all is.

Agentic campaigns changed this by running multiple recommendation strategies in parallel, doing the segmentation work automatically.

MandM tested five homepage recommendation strategies simultaneously. Rather than splitting traffic across two variants and waiting two weeks per test, Made With Intent's optimisation agent tested all five, across 157 unique segment combinations, and allocated each visitor to the strategy most likely to drive Ollie's defined commercial goal; revenue per user.

The result was a 2% increase in revenue per user. But the more valuable output was just how granular MandM could get. Not just "strategy X wins." For MandM it was strategy X wins for loyal mobile customers, strategy Y wins for new desktop visitors, and for your most engaged customers, showing recommendations at all might be the wrong call.

Let that sink in: showing recommendations at all might be the wrong call.

On the PDP, where new customers arriving from paid search land and recommendations have some of the most direct commercial impact, MandM ran a similar test with five strategies including a "hidden" variant (no recommendations shown at all). The result: 4% increase in revenue per user, from 130 unique segment combinations tested.

On the webinar, Colin Spooner, Principal Value Consultant, at Made With Intent, describes what this looks like inside the platform:

"There's kind of no winners or losers anymore. It just is the best experience to give that visitor at the right time."

Sometimes, doing nothing is the best strategy

In MandM's PDP test, 14% of visitors were allocated to seeing no recommendations at all, and for that segment, it was the highest-performing option.

That segment, as Ollie describes it, is your most loyal customers. The ones who know the site, know what they want, and don't need or want a carousel of "You might also like" items interrupting their path to purchase.

Ollie says: "For a certain subset of customers, your very loyal customers, the ones that know the site, they know what they want, removing recommendations is actually beneficial. Sometimes it can be a bit of a loop for a customer."

The PDP recommendation loop is a real issue. A customer lands on a PDP, clicks a recommendation to another PDP, clicks another, and ends up in a browse abandonment spiral that started as a purchase intent session. It's almost like you're giving them too much to navigate through.

Removing the recommendations breaks the loop and lets the customer do what they came to do.

This can be uncomfortable for personalisation teams to hear whose KPI is coverage, ensuring every visitor gets served something. But it reflects a more mature way of thinking about personalisation: not "show more" but "show what's right, when it's right."

For some customers in some moments, the right thing is nothing.

What MandM's results point to

MandM's results across two distinct experiments expose the same fundamental flaw in how most onsite experiences are built. They're designed to give every customer an answer at predefined moments, when real impact is about giving each customer the right answer for their specific moment.

Rule-based email capture fires at step three of the journey regardless of whether the customer is engaged or about to leave. Sequential recommendation testing finds one winning strategy for one segment, then starts over.

Ollie used intent signals and agentic campaigns both push against that. One changes when you fire an experience based on real-time behavioural signals. The other changes what you show based on continuous, parallel testing across hundreds of segment combinations. The outcome, in both cases, is the same — fewer experiences wasted on the wrong customer in the wrong moment.

For MandM, the next step is applying agentic testing to placement-level messaging. So, buy now pay later, delivery propositions, app downloads, and letting our agent match messages to customer segments in real time across the same placements.

If you enjoyed this blog post, why don't you watch our Intent Live series over on YouTube? Failing that, we're going to be running these sessions frequently, so you can sign up here for our next session.

While we're doing CTAs, here's another one: If this has piqued your interest, why don't you book a demo with our team?

Most brands only respond to cart abandonment after it's happened. That's like fitting a smoke alarm and calling it fire prevention.

Seven out of ten online shopping carts are abandoned before checkout. That number hasn't meaningfully shifted in a decade.

Not because the industry hasn't tried. Cart abandonment emails are everywhere. Retargeting ads chase shoppers across the internet for days. Brands have invested millions in user recovery, the machinery that kicks in after someone walks away.

And it works. Abandoned cart emails recover between 3% and 5% of lost baskets on a good day. But the thing about existing cart abandonment strategies is that they treat the problem after it has already happened.

What if there was a way to intervene before someone dumps their cart of goods? How would you get this visibility? In this blog post, we explore cart abandonment, discuss existing strategies and how new technology can provide timely interventions to stop basket abandonment in the moment.

What is cart abandonment? And why does it still matter?

Cart abandonment happens when a shopper adds items to their basket but leaves the site before completing a purchase. The global cart abandonment rate sits at roughly 70%, according to Baymard Institute's aggregated research across 49 studies — and it's been stubbornly consistent for years. For UK ecommerce brands doing millions in online revenue, that 70% represents an enormous amount of commercial value slipping away every single day.

The industry has treated this as a recovery problem. It's actually a visibility problem. Brands can't see why or, specifically, when shoppers hesitate, so they can't respond until it's too late.

The abandoned cart email: Essential, but not enough on its own

A well-built abandonment flow is one of the highest-ROI programmes in ecommerce, and the brands doing it well deserve credit for that.

But the returns are diminishing, and the reason is structural, not tactical.

A decade ago, a well-timed abandoned cart email felt personal. Now it's expected. Shoppers know the email is coming. Some even use it as a strategy: abandon the basket, wait for the discount code, complete the purchase at a lower price. If you always send these emails to every cart abandonment, you've trained consumers to do this.

The average abandoned cart email open rate is around 40%. That sounds impressive until you realise that fewer than half of those opens result in a click, and fewer than half of those clicks result in a purchase. You're recovering a fraction of a fraction.

The email programme isn't the problem. The problem is that it's the only thing working on abandonment. Everything that happens before the email. The entire session where the shopper was actually on your site is a gap.

Why shoppers abandon baskets (and why most brands get the diagnosis wrong)

Ask any ecommerce team why shoppers abandon baskets and you'll hear the same list: unexpected shipping costs, complicated checkout, required account creation, security concerns. Baymard Institute's checkout usability research has documented these friction points extensively, and they're real.

But they're also the easy answers. The ones that show up in surveys and exit polls because they're concrete and simple to articulate.

The harder truth is that most basket abandonment isn't caused by a single friction point. It's caused by unresolved hesitation.

A shopper adds something to their basket. They're interested, clearly — but they're not convinced. 

Maybe they're comparing prices elsewhere. Maybe they're not sure about sizing. Maybe they need to check with a partner. Maybe they're thinking about considered purchase and this is visit two of five. Maybe they’re just building a wishlist, and were never going to buy anyway?

None of these shoppers have a checkout problem. They have a confidence problem. And no amount of checkout optimisation or recovery email will fix that — because by the time the checkout loads or the email arrives, the moment has passed.

The real gap: what happens during the session

Let’s try a thought experiment. Let’s categorise ecommerce into black and white terms. On one end: acquisition, getting shoppers to the site. On the other: recovery, trying to win them back after they've left. The bit in the middle, the actual shopping session, is where you have the least visibility and the fewest tools at your disposal,.

Think about what a good shop assistant does in a physical store. They don't wait until you've put something down and walked out, then chase you into the car park with a voucher. They read the room. They notice when you're browsing versus when you're comparing. They step in when you look uncertain and step back when you're clearly decided.

Online, we do the opposite. We show everyone the same experience — same pop-ups, same messaging, same urgency banners — regardless of whether they arrived 10 seconds ago or have been comparing products across three sessions over two weeks. Then, when they leave, we send the email.

The gap isn't in recovery. It's in the session itself. The question isn't "how do we get them back?" it's "why didn't we respond to what they were telling us while they were still here?"

What in-session card abandonment intervention actually looks like

In-session intervention means responding to shopper behaviour during the visit, not after it. But (and this is the critical part) it doesn't mean bombarding people with more pop-ups and discount codes.

The problem with most "onsite intervention" is that it's based on static rules, and doesn’t interject at the moment. It’s usually after the moment has passed. Show a pop-up after 30 seconds. Trigger an exit-intent overlay when the cursor moves toward the tab. Offer 10% off to everyone who has items in their basket.

These rules treat every shopper the same. A first-time visitor browsing casually gets the same intervention as a returning visitor who's viewed the same product four times and is clearly ready to buy. That's not intervention. That's bad manners.

Real in-session intervention requires knowing where a shopper is in their buying journey — right now, in this session, and responding appropriately.

For a visitor who's browsing early in their journey: don't push. Show them content. Help them discover products. Surface reviews, comparisons, and reasons to believe. The worst thing you can do is ask for a commitment before they’re ready.

For a visitor who's been comparing across multiple sessions and has returned to a specific product: they don't need a discount. They need reassurance. Delivery information. Stock availability. Social proof that others bought and loved this item. Remove the doubt, and they'll convert without a price incentive.

For a visitor showing every signal of purchase intent but hesitating at the basket: now a small nudge might help. Free delivery. A modest discount. A reminder that the item is selling fast. But even here, the intervention should match the hesitation, not just throw money at it.

This requires intent data, the ability to model behavioural signals within a session and predict where a shopper is in their buying journey before they abandon.

Full disclosure: this is what we do at Made With Intent. But the principle holds regardless of how you implement it. If you can see where a visitor is in their buying journey, you can respond before they leave. Whether you build that capability internally, stitch it together from your existing stack, or use a dedicated tool, the strategic point is the same.

The economics of basket abandonment prevention  versus recovery

Let's put some numbers on this.

A mid-market ecommerce brand with 500,000 monthly sessions, a 3% conversion rate, and a £75 average order value generates roughly £1.1M per month. At a 70% cart abandonment rate, shoppers are adding items to baskets in around 50,000 sessions but only completing 15,000 of those purchases. Roughly 35,000 baskets are abandoned every month.

A strong abandoned cart email programme recovers 3–5% of those, say 1,400 orders, worth around £105,000 per month. That's meaningful revenue, and it should absolutely keep running.

Now consider what happens if you can prevent even a small percentage of those 35,000 abandonments from happening in the first place. A 5% reduction in abandonment just by delivering better in-session experience would save 1,750 baskets. At £75 AOV, that's £131,000 per month in revenue that was never lost and never needed recovering.

Prevention requires investment, tooling, configuration, and ongoing optimisation. It isn't free. But here's where the margin argument gets interesting.

Recovery emails frequently include a 10% discount as the incentive. On £105,000 of recovered revenue, that's £10,500 in margin given away every month. Not because every one of those 1,400 shoppers needed a discount to convert, but because, without real-time intent data, you have no way to know which ones did.

This is exactly the problem Better Bathrooms faced. Without visibility into visitor intent, they were stuck in a choice most ecommerce teams know well: discount broadly and erode margin, or do nothing and accept the exit rate. Neither option was good enough. 

By activating in-session intent signals to target discounts only at visitors who had built purchase intent but were showing signs of dropping off, they broke out of that trade-off entirely, lifting conversion rate by 26% while protecting the margin they'd previously been leaking. The discount didn't change. The targeting did.

Over a year, that margin difference compounds significantly, often enough to fund the prevention capability several times over.

The two approaches aren't in competition. They're complementary layers. But most brands have invested heavily in recovery and barely at all in prevention. The opportunity is in rebalancing that investment.

Why brands haven't done this already

If in-session intervention is so effective, why isn't everyone doing it?

Because until recently, brands couldn't see what was happening during the session in a meaningful way.

Most ecommerce teams work with two types of visitor data. Historical data tells you what someone did last time, with things like their purchase history, their email engagement, their lifetime value segment. Page-level data tells you what page they're on right now. Neither tells you where they are in their buying journey in this session.

Are they browsing or buying? Are they comparing options or ready to commit? How likely are they to buy or abandon their purchase ?

Without answers to these questions, every visitor with items in their basket looks the same. You can't intervene differently because you can't see differently. So you wait until they leave, and you send the email.

Intent data changes this equation. By modelling hundreds of behavioural signals within a session — scroll depth, navigation patterns, time on page, return visit frequency, basket interaction, comparison behaviour — it becomes possible to predict where a shopper is in their buying journey before they abandon. Not after.

What this means for your abandonment strategy

If your entire cart abandonment strategy is a post-session email flow, you've built one layer of a two-layer system. The email works. The question is what sits alongside it.

Here's what an intent-based abandonment strategy looks like:

During the session: Use intent signals to identify visitors who are showing signs of hesitation. Respond with the right experience, reassurance for the nearly-convinced, content for the still-exploring, and targeted incentives only where they're genuinely needed.

At the point of exit: If a visitor does move to leave, your exit-intent experience should be informed by their session behaviour, not a one-size-fits-all overlay. A returning high-intent visitor doesn't need 10% off. They need a reason to buy now.

After the session: Continue running your abandoned cart email programme. But let it be the safety net beneath a more complete strategy, not the whole strategy. And personalise those emails with intent data from the session — a shopper who was comparing options needs different messaging than one who got to the payment page and stopped. For instance, if someone has multiple items in their basket, which item did they have a real preference for if any?

In addition, because you’ll be sending fewer cart abandonment emails, and only sending them to the visitors with appropriate intent, your number of sends and unsubscribes will naturally go down.

Across sessions: Recognise returning visitors who previously abandoned baskets. Their second visit is the highest-value moment in the entire journey — they came back because they're still interested. Don't waste it by showing them the same generic experience they saw last time.

The 70% isn't going away. Your response to it should evolve.

Cart abandonment isn't a problem to solve. A 70% abandonment rate reflects the reality of how people shop online — browsing, comparing, considering, returning, and eventually buying. Fighting that reality is a losing game.

What you can change is how completely you respond to it. Most brands have built strong recovery programmes. That's the foundation. The next step is building the capability that sits before recovery — the in-session layer that catches hesitation while it's still happening, responds to what each visitor actually needs, and prevents a portion of those abandonments from ever reaching the email queue.

The brands that build both layers don't just reduce their abandonment rate. They reduce their dependency on discounts to recover lost revenue. They protect their margins. And they build a better experience for their shoppers — one that responds to what visitors need in the moment, rather than chasing them after the moment has passed.

Your abandoned cart email is the safety net. It should stay. But the real opportunity sits earlier — in the session, in the signals, in the moments where the right experience could have kept that shopper moving forward.

Made With Intent helps ecommerce brands see where every visitor is in their buying journey and respond in the moments that matter, reducing cart abandonment. If you want to add the in-session layer to your abandonment strategy, book a demo.