2022 marks the culmination of two years’ investment in behavioural science, machine learning and deep segmentation at home improvement brand Wickes.
Praised for its innovative approach, the retailer swept the board at the Marketing Week Awards in November, winning the awards for Retail and Ecommerce, Best Use of Segmentation and the coveted Grand Prix.
The origins of Wickes’ success can be traced back to the lockdown-fuelled DIY boom of 2020. Trapped at home, UK consumers decided it was time for a revamp, meaning Wickes was in high demand.
The digital business grew tenfold, click-and-collect volumes soared by 400% and home delivery sales rose by 100%. Wickes’ customer base doubled, causing a wealth of customer data to pour into the business over a short space of time.
The need to turn data into insight was first voiced by chief marketing and digital officer Gary Kibble before lockdown even began.
Kibble, who joined the DIY retailer in January 2020 from his previous role as marketing director at Argos, sees Wickes as having a “competitive advantage” based on being “digitally led and service enabled.” Just over two thirds of all sales start online, 98% of which are “touched” by a store in some capacity.
The £7m [incremental revenue] is tip of the iceberg stuff, which is exciting.
Gary Kibble, Wickes
That said, Kibble insists data without insight is largely irrelevant. This plays into his mantra: data to insight to action to create improved outcomes. The sheer amount of data “amplified” by Covid made executing this mantra more relevant than ever.
The process began by digging into three years of historical data, working with partners Team ITG and data science specialist Emerald Thinking to identify patterns in browsing and buying behaviour associated with specific DIY “missions”.
The result was the Mission Motivation Engine, a machine learning model incorporating online and in-store transactional, search, browsing, engagement and third-party data, with insight into how consumers click on social, display and website content.
“You need billions of rows of data to look for the common products purchased within a mission. The minute you start to go back over three years, you start to get a real sense of the sorts of products consumers buy or browse within a mission,” Kibble explains.
“That allows you to group up into common missions. You can only do that if you have loads of data. You couldn’t do that if you said: ‘I’m going to look back over the past month.’ You wouldn’t have enough patterns in the data to draw the conclusions you are able to draw when you’re combining browsing and buying over three years.”
Once the team started to identify the commonly bought and browsed products, they were able to get a sense of consumer behaviour and assign them a mission. Kibble uses the example of a room renovation as a type of mission.
“If you wake up on Saturday and think you’re going to paint your room, you can’t just paint it. You need paint, brushes, sanding blocks and paper, masking tape, dust sheets, white spirit. You therefore can build up a mission, which is a full end-to-end of all the products you need,” he explains.
“What we saw in the data is too low a proportion of customers were buying the full mission from us, which led to us starting to say: ‘What are the missions that exist in our business for customers? What proportion of customers buy what proportion of the mission? How do we start to build content and communication that helps them engage with us more frequently on the end-to-end mission?’”
The data showed, for example, consumers associated Wickes with specific products, rather than their whole project, with tradespeople tending to shop from six categories or less and supplement with rival brands.
With the machine learning model in place, Wickes was able to feed this customer insight into a new segmentation model.
At the highest level is the mission segmentation, below which sit various sub-segments including aptitude. The AI-powered engine identified 10 DIY/showroom missions, seven trade professional missions, seven motivational trade segments, 11 TradePro programmes and 10 DIY/showroom programmes.
A customer embarking on a room renovation might be a novice DIYer or advanced tradesperson, meaning while the overall mission is the same, the communication strategy should differ.
“Within the site we created and tagged every single piece of content. So, for example we identify you’re on a room renovation mission. I start communicating to you about the thing that is going to help you and guide you most,” says Kibble.
“However, if you’re an expert DIYer and your aptitude is very high, I would serve up different content, potentially different products. Rather than serving you up own brand paint, I might serve you the best – as in good, better, best – and I might serve you different content.”
The team are currently developing a filter focused on attitudes to sustainability, based on a UK-wide study identifying five different segments. This added filter will be layered over the mission and aptitude filters to give a “meaningful” read on the kinds of products a consumer might be looking for.
“The only challenge you always have with data is when is the juice not worth the squeeze? Sometimes you can create so many sub-segments you create so much work that the return you generate off that investment and time doesn’t pay back,” Kibble notes.
“The trick is knowing at what point we’ve gone low enough in our segment and we don’t need to go lower. The maximum ROI will be delivered at that level.”
Getting the wider company to embrace this AI-powered approach to segmentation was relatively straightforward. Kibble believes it’s marketing’s responsibility to be the company’s conscience when it comes to innovation.
“As marketers we always have to put ourselves out there with new technologies and platforms, whether that be the metaverse, or in this example machine learning and AI. You have to put yourself out there and you don’t always get it right,” he notes.
“This was something that was always going to work. You’ve only got to put yourself in the customer’s shoes to understand how you interact with retail brands in the home improvement space to know if I was shopping for a room renovation mission, I probably wouldn’t go to one retailer and buy that whole mission.”
Understanding the commercial team’s priorities also helped. Kibble worked in the commercial function for the first decade of his career, serving as business unit director for books at WH Smith. He acknowledges commercial teams often believe the best route to engage customers is volume, preferring an email to reach 5 million customers over a smaller group who are highly engaged.
“If you had a commercial mindset that said: ‘Tell as many people as you can about my promotion, activity and product,’ move into a world where you’re only going to talk to the people where it matters most, that’s a big mindset shift,” Kibble explains.
The way to convince the commercial team to make this shift was to run the new activity alongside the ‘business as usual’ programme, presenting the numbers which proved targeting customers via missions was working.
“If you’re a commercial individual money talks, so that’s all it takes to convince them it’s a good thing,” he adds.
Driving strategic growth
Based on the new segmentation strategy, Wickes revamped its communications. The team decided to ditch product-focused comms in favour of motivational messages demonstrating how Wickes can help customers achieve their goals.
Communications are triggered when a purchase is ‘spotted’ by the model and timed to match the length of the project. The retailer combines email, app push, social and landing page channels to guide consumers through their DIY project. The tone shifted from talk of sales to friendly hints and tips.
Based on insight that tradespeople prefer to shop on Sunday night and Monday morning to get organised for the week, the marketers developed weekly communications under ‘The Week Ahead’ banner, adopting a new tone of voice and personalised to-do list designed to position Wickes as the trade’s default supplier.
It was crucial for Kibble the approach stretched far beyond an email communication programme on owned channels, to platforms such as streaming services and targeted radio messages.
“So, working with Carat our media buying agency to say: ‘Here’s a segment of customers on a room renovation mission, therefore where we see them showing up on More 4 we want to put the relevant content in front of them,’” he explains.
The trick is knowing at what point we’ve gone low enough in our segment and we don’t need to go lower.
Gary Kibble, Wickes
Having measurement in place to give the wider team confidence the strategy is working has proved crucial. The overarching KPI the marketers focus on is revenue per customer. The team also analyse whether Wickes is increasing the proportion of each mission being purchased by customers.
A revamped welcome programme for new customers has driven an improvement in the volume of customers progressing from a first to second purchase, success Kibble attributes to his team’s ability to use data to “build out a narrative for customers.”
“The last area of success has been our ‘win-back’ programme for those dormant customers, so how we’ve used historical data to bring them back into the brand,” he explains.
“There’s four key metrics at that high level, but ultimately it all flows through to that one metric which is revenue per customer. We had a full suite of KPIs, but that’s the one we really obsess about.”
The marketers are currently building a suite of customer personas through the Mission Motivation Engine. They have also introduced monthly ‘Customer Closeness’ sessions, inviting real customers to discuss key issues such as the impact of the cost of living on the building trade. People from the business can join the sessions to ask questions, or simply watch and listen.
In his nearly three years at Wickes, Kibble feels most proud of shining a “high beam spotlight on the customer”. The minute a business starts to focus on the customer, he believes people begin to act and think differently.
“We’ve brought the customer to life and any business that does that for me is going to be successful. It’s not rocket science, just bring the customer into your business,” he advises.
From a financial perspective, the investment in machine learning and segmentation is clear. Wickes generated more than £7m in incremental revenue within six months of implementing the Mission Motivation Engine.
CEO David Wood sees the engine as a “differentiating IP and a strategic growth lever”. The AI-powered model has been part of the brand’s investor pack since it was introduced and Wood is frequently asked for updates on its progress by analysts, describing it as “real catalyst” for growth.
Kibble appreciates the fact Wood is an ex-marketer, who served as CMO at Tesco, marketing director at Mondelez and within the marketing function at Unilever.
“We speak the same language. We just get each other and you can see the value that brings,” says Kibble. “That isn’t always true with other CEOs I’ve worked with who maybe had more of a CFO background. He understands the role marketing plays around the top table.”
With backing at the highest level for the AI-powered strategy to continue, the Wickes marketing boss believes his team are only “in the foothills” when it comes to commercial and customer outcomes, as well as the ability to scale.
There are multiple Mission Motivation Engine programmes set against Wickes’ trade, DIY and showroom businesses – the latter relating to installed kitchens and bathrooms. Kibble estimates the team are just 30% of the way through these programmes, on average.
“Therefore, we’ve got 70% of the programme still to come,” he explains. “We think this will give us a material competitive advantage from a commercial and customer perspective. The £7m [incremental revenue] is tip of the iceberg stuff, which is exciting.”