Why Stuart Machin is ‘Positively Dissatisfied’ with The Personalization at Marks & Spencer10 min read
Table of Contents
As I gear up for my December vacation in an exotic locale, my quest for the perfect winter wardrobe and party wear led me down the internet rabbit hole. Amidst discovering a few enticing winter apparel pieces, one news article grabbed my attention.
Stuart Machin, the CEO of Marks and Spencer, was heralding the news of surging online sales and robust profit margins.
However, what intrigued me was his candid confession of being ‘positively dissatisfied’ with the company’s personalization endeavors.
With a reported 4.6% surge in online sales and a 9% boost in margin entirely attributable to a reduction in discounted items and increased revenue, Machin believes improved personalization could have further elevated their profits.
A paradox in success, and a revelation that sparked my curiosity about the untold story behind the scenes.
M&S Personalization Strategy – A Backstory
Back in 2020 when Marks and Spencer initially embarked on developing their personalization approach, they honed in on three key strategies:
- Making a strong business case for personalization among shareholders by clearly identifying value streams in CRM, digital marketing, and digital experiences, aligning team efforts with revenue and value streams.
- Developing a channel-agnostic single customer view for personalization to cover diverse touch-points and avoid siloed customer analysis by teams.
- Balancing metrics that help understand the impact of their actions at an individual use case level by shifting from last-click revenue to incremental revenue.
The results? A remarkable 1.5x increase in revenue the first year followed by a 2x climb in the next.
However, every success story has its nuances, and upon a closer examination of their site, I uncovered areas where the personalization strategy could have been further refined. Let’s delve into these nuances for a comprehensive understanding.
Muddled Data – The Tagging Problem
For any company, data serves as the lifeblood, and this holds particularly true for e-commerce platforms, given the vast array of products they manage.
As a shopper, my journey typically begins with a search on Google or directly on an e-commerce website. Regardless of the entry point, my expectation remains consistent — a relevant, personalized, and diverse collection of products to choose from.
For a smooth shopping experience, e-commerce sites need clear product descriptions and accurate tags.
Confusing data – poorly organized catalog information and an unstructured taxonomy for products, make it hard to find products and compromise a personalized shopping experience.
Sadly this is something that I noticed on the M&S site. When searching for a ‘black party dress,’ I was presented with just two results – really? That seemed off.
I refined my search to simply ‘party dress,’ and the autocomplete feature completed it as “Party dress for women” – a nice touch.
However, the disappointment resurfaced when the result page included dresses for kids in response to the ‘party dress for women’ search.
Oh there, when I looked for winter wear, the results kindly suggested an “Active zip-through swimsuit” and “Suede slippers with Freshfeet”. Because, you know, frosty weather calls for a swimsuit stroll and toasty slippers.
Nice try, algorithm, but I’ll stick with my winter coat, thanks.
Now that’s what I call a data problem.
Proper product tagging goes beyond mere surface-level information. It involves crafting detailed tags and product metadata, derived from both visual and textual inputs. It enables retailers to clean, structure, and label product data efficiently.
With relevant information gleaned and attached to the products, it is easier for the AI to pick and deliver a contextually relevant result for a given search query.
Because at the core of it, Clean Product Data + Customer Data + AI Personalization = Retail Success.
Pro-tip: Showing search results that are not only relevant but preferably within the first scroll has shown higher clicks and in turn increased conversions among customers. So don’t forget personalized sorting.
Dynamic 1:1 Personalization
67% of consumers think it’s important for brands to automatically adjust their content to reflect the current context – Adobe
Not all shoppers are the same, and a one-size-fits-all approach falls short. You want your customers not just scrolling but adding goodies to their carts, right?
Think about it – each shopper has their groove. It’s not just about selling; it’s about building a connection. Real connections mean loyal customers and loyal customers mean a sky-high lifetime value (CLTV).
Not everyone’s shopping for themselves. Knowing what your customers want before they even realize it themselves? That’s the game-changer.
Imagine their surprise when your site adjusts in real-time, showing them exactly what they didn’t know they needed.
The magic of turning the static vibes of 2020 into dynamic 1:1 personalization – boosts customer satisfaction, increases the chances of purchase, and helps products get noticed.
What immediately grabbed my attention was the M&S category page. Despite my looking around the site for party wear and winter wear the results seemed immune to any dynamic personalization based on my clicks and scrolls.
Navigating to the ‘Women’ > ‘Partywear’ category, I found recommendations that checked the party wear box but were nowhere near my personal preferences.
At the bottom of the page, a small ‘Recently Viewed’ widget hinted that the website was capturing my browsing history. Still, alas, the AI struggled to translate that into spot-on recommendations.
This is exactly where 1:1 Dynamic Personalization can step in and make things right.
Vue.ai’s personalization solution can dynamically personalize recommendations that are shown on the category page based on user behavior like clicks, scrolls, and intent.
It is achieved by creating individual shopper profiles based on customer data and their behavior on-site combined with product data. The AI understands shoppers’ intent with every click, search, and action on the site, adjusting personalization in real time.
Styling and Similar Recommendations
Are you aware of the primary factor that leads to high bounce rates for e-commerce shoppers? It’s the inability to discern shopper intent and insufficient visibility of relevant information.
In an analytics study, conducted by the Vue.ai team, over 4% of all add-to-carts (ATCs) on a site came from Visually Similar Product Recommendations. Showing products of high relevance and matching user intent, increases customer satisfaction and leads to higher sales conversion.
Similar recommendations reduce shopper abandonment by surfacing products that users might have otherwise missed exploring. But making sure where to put this recommendation and what you should pair it with is also crucial.
While exploring M&S, although I did find this option on their product page, I found there were numerous personalization widgets stacked one after another.
Take, for instance, the product page for the ‘Sequin Slogan Crew Neck Jumper,’ which had about three distinct recommendation widgets tucked away below the page.
- Style it with
- You may also like
- Why not try
It seemed like a cascade of choices that could potentially overwhelm shoppers. The key lies in testing and finding the sweet spot for optimal personalization and enhanced user experience.
In my opinion, there’s an overload of options for shoppers to navigate. The consecutive presentation of these choices risks inducing decision fatigue.
An optimal personalization setup would entail placing the most relevant “Style it with” immediately alongside the product, with “You may also like” positioned beneath it.
Determining the ideal placement for these personalization modules is best achieved through systematic testing.
Vue.ai takes the lead here. Its algorithms, steering visually similar products, decipher user intent with a simple click, breaking down the visual style attributes of any given product.
This contrasts with a visual search that requires the algorithm to precisely understand the customer’s search. Vue.ai, on the other hand, integrates these visual affinities into a personalized Style Profile for each shopper.
How vue.ai builds 360-degree customer profiles based on preferences and behavior
With Vue.ai’s A/B testing tools, retailers can seamlessly cross-experiment with personalization widgets and placements. This facilitates the optimization of recommendation results across the entire website and on individual product pages.
Increase upsell with Alternative recommendations and Personalized Email
Acquiring a new customer can be a pricey endeavor—five times more expensive than retaining an existing one, as per LinkedIn insights. Therefore, it’s prudent not to overlook the efforts of shoppers who’ve actively visited your store and explored your products.
Every e-commerce entity strives to boost its Average Order Value (AOV) and expand sales, presenting a realm of possibilities.
While navigating the cart page of M&S, I spotted an opportunity – the addition of a “Why not try” widget or perhaps dubbing it “Pair it with” or “Style it with” could work wonders.
Leveraging the cart page for effective upselling is a strategic move. Incorporating personalization options, like offering “complimentary product recommendations” during the first email sign-ups, is a valuable approach.
Understanding that users have explicitly shown interest in your product allows you to capitalize on that interest, guiding them toward a purchase with targeted recommendations.
So, wrapping up my deep dive into Marks and Spencer’s personalization world – it’s been a rollercoaster of insights.
Lessons learned: Product tagging is a big deal. Muddled data is like a foggy road; you can’t find what you’re looking for. And let’s not forget the category and result pages – they felt a bit like a static party in a dynamic world.
In the ever-shifting e-commerce game, it’s not just about hitting a target once and calling it a day. It’s about the constant hustle, the tweaks, the ‘Aha!’ moments. Clean data, understanding your customers, and letting AI do its thing – that’s the retail magic formula.
Marks and Spencer might have some fine-tuning ahead, but hey, who doesn’t? The journey never ends; it’s a vibe, a dance of optimization. So, here’s to retailers everywhere – keep it real, keep it dynamic, and make every click count.