Personalisation chiefs from leading global retailers detail the foundations they've been building to ensure personalisation succeeds and can be done at scale
Personalisation is one of those buzzwords in marketing that means something different to every brand you talk to. But there are commonalities when it comes to building the foundations for an organisational personalisation strategy that can help set brands up for success.
During the recent Adobe Summit, personalisation strategy leaders from two dominant global retailers – M&S and Victoria’s Secret – shared key steps they’d taken to build out respective personalisation strategies, from measurement frameworks and initial business cases to how to find the right areas of personalisation first, goal setting, cultural buy-in and more.
Here, we present several of their useful learnings and must-take steps.
1. Build a realistic, compelling business case
The most important step is being able to articulate a business case for personalisation in the first place. M&S head of growth and personalisation, Alex Williams, said his team used widely available benchmarks in order to understand important value streams and how they would grow the revenue line short to long term.
“We identified clear value streams aligned to how our company thought about our core activities to customers: Offers, CRM and digital marketing and digital experiences,” he explained. “Loads of work was then done to assess what uplifts are. We used industry reports such as Adobe’s Retail Trends Report, and heavily relied on BCG Gamma and its work in this space. We had a good understanding of personalised offers internally, so it was easier to project what an expected uplift could be.
“In CRM/digital marketing, the focus was on specific use cases. For example, suppression of highest value customers, who tend to visit the website or come in-store anyway, plus building new trade-driven triggers into our email program.”
For digital experiences, including the M&S app and website, Williams’ team projected out a share of revenue uplift expected based on benchmarks available.
“We realised when we think about this from cost perspective that we were often displacing cost of agencies internally running CRM program or personalised offers for us,” he said. “So for value streams through, we focused on cost structure of the team and a potential revenue stream where the ROI of that made sense and we could manage on a quarterly or annual basis. It was about aligning team to revenue and value stream.”
So far, so good. But after pulling this together and taking it to stakeholders initially, Williams said many were unconvinced on deliverables. “They thought it was unrealistic to expect building from a base on zero,” he said.
“We went back through and pulled several cost-saving use cases forward to gain more confidence, removed the steep hockey stick we had in the plan, as we were putting in infrastructure investment early and seeing steep returns later, and phased that out over time. We got to a point where we felt comfortable to move forward with the plan and lots of healthy haircuts here and there.”
In practice, M&S blitzed its revenue target by 1.5x in the first year and 2x in the second year. But this revision and balance of short and long-term projection proved critical in getting overall buy-in early on, Williams said.
At Victoria’s Secret, the personalisation project started with exploration, lots of user research and competitive landscape assessment. This led to its ‘north star’ development, where teams aligned enterprise-wide on a definition of personalisation and defined eight customer-facing outcomes and built consensus cross-functionally within the company, said head of digital products, Jenna Brunner.
2. Determine a comprehensive measurement framework
According to Williams, creating a suitable personalisation measurement framework is the most undervalued and therefore underinvested in capability in an organisation.
“Without it, it’s very hard to understand what works,” he said. “We tried to establish a framework that comprised a set of balancing metrics – things that work alongside one another and where the balance of both gets you to a good outcome. This is opposed to a single measurement, which can drive you headlong down a route without fully understanding the implications.”
In M&S’s case, the big shift was from last-click revenue to incremental revenue. “This allowed us to understand at an individual use-case level whether the changes we were making were having the uplift, then measure and bank that with the finance teams,” Williams explained.
In addition, M&S tracks the number of personalised interactions on Web and app via Adobe Analytics and number of emails automated to understand scale of personalisation efforts.
“We continued to report on short-term last-click revenue, because in introducing these new incremental measurements naturally feels uncomfortable in initial stage and it gives us some way of connecting through in the framework,” Williams said.
Over at Victoria’s Secret, measuring personalisation’s impact in a more holistic way and through language the business related to was also critical. Brunner said this was done through a “personalisation backtest”.
“We didn’t have a way to measure the total incremental impact of all our personalisation activities. We could only measure each test. The question was how to sum all these individual tests up to understand true impact,” she said. “We needed a reliable way to measure incremental of total ecommerce personalisation efforts.
“What we did was leverage our content management testing solution and hold back 5 per cent of all site visitors from receiving any personalised experiences. We analysed KPIs of this control group versus test group, then ran this test for 12+ months and reported on results to leadership team on monthly basis.
“We saw double-digit incremental growth over a 12-month period – a 15 per cent incremental lift in revenue for our digital business by using personalisation capabilities.”
Victoria’s Secret has since stopped this test to realise the benefits of personalisation across its total digital audiences. But as Brunner put it, “by doing this, we drove lot of confidence among enterprise to not need this holistic analysis approach”.
3. Know what your single customer view really consists of
Having the right kind of customer view was another critical component in how M&S has fuelled its personalisation strategy.
“Single customer view is something that sounds good, and if you ask your teams, they will respond they have one. Unfortunately, most are a siloed view of customer behaviour,” Williams claimed.
“Typically, we see digital marketing creating their own set of data around marketing response, and ecommerce around all behaviour on the website from browse to transactions. Those retailers who have a store base might have those connected in, but you end up with two different sets of experiences you are delivering, leaving a load of value on the table.
“From the beginning, we started building towards a personalisation capability that had a single customer view at its centre that was channel agnostic and could incorporate any future business line or touchpoint into it to create as much scalability as possible.”
To do this, M&S has three layers: A customer data layer, which stitches together and incorporates product, transactional, behavioural and customer data; an intelligence layer, which houses all data science and machine learning models; and an orchestration layer to control all those experiences.
Read more: 9 lessons in tackling customer retention and a single customer view at Carsales
4. Find differentiated brand use cases for personalisation
When it came to working out what to personalise, Brunner and Williams pointed to a raft of out-of-the box capabilities they initially utilised thanks to investments in technology platforms. What was also critical was to identify specific use cases that provided differentiation in market and met with specific business objectives.
At Victoria’s Secret, one of these was what Brunner called 'automated matchbacks'. This relates to a section on its product page that highlights matching items to a product viewed by a visitor.
“One of our biggest goals as a business is to sell the total outfit. The problem identified was internal teams were manually assigning matching items, not based on any customer behavioural data,” Brunner said. “The outcome was to see matching items to primary product powered solely by Adobe Target. Today, visitors see matching items other customers have bought with it. That’s set up within our Target activity and swatch bar to match on same swatch or colour as primary product. So as she clicks, the items will change to match her newly selected colour.”
This resulted in single digit lifts in revenue per visitor metric. Brunner said her team continues to iterate, running more than 30 tests to see what drives highest growth here. Another benefit was eliminating manual workload with a more performance-driven solution.
A second win for Victoria’s Secret was personalising content through ‘auto-allocate’. The problem identified was the same content was displayed to all site visitors on most visited landing pages regardless of whether they were a new or existing customer. In response, Brunner’s team set out to increase engagement and revenue from new customers coming to this page with more personalised educational content that could walk a customer through how to buy a bra online. Existing customers see more fashion product content.
“We saw significant growth by displaying more educational content to this new customer segment: On average, a mid-teen lift in revenue per visit for new customers,” Brunner said. “We also utilised auto-allocate to determine the right cues as we didn’t know. This took the guessing out of it for us.”
A third win was to digitise seasonal customer rewards using Adobe Campaign and offers. Previously, customers received a physical card in packages and had to save the unique 11-digit code for several weeks then type it in online when it was time to redeem the offer. To make it easier to redeem and display these offers to customers, Victoria’s Secret sent unique email coupon using Adobe Campaign.
“When that customer returns online, we can display an offer sheet as a reminder and allow customers to apply the offer once it’s in their offer drawer,” Brunner said. “We saw single-digit revenue lift per visit, providing a conversion bump.”
Both out-of-the-box and internally built recommendations at different parts of the customer were very important to M&S too. “We used Adobe Target on M&S.com to get quick wins, playing with features that fed into those algorithms so we could optimise for different outcomes,” Williams said.
“We also looked internally, including at our outfit recommendation capability. This showcased where there was incremental value to be had from us building our own capability – in this case, a manual outfitting service we had and all the data related to that. This showed how we could create flexibility once we built this use case to push into multiple other channels and generate scale much more efficiently.”
5. Show the value and path towards scale
Williams and his team mapped out the first year of its program using Miro. In the centre were all the data science and models, then around the outside were use cases and channels they sat in, such as app, push, email, offers and paid marketing, with each showing value created. This was shared with the board plus stakeholders.
“The interesting thing about this approach is although you need a new model for each use case in the beginning, as you build these out, you can use these models in multiple different use cases and scale much more quickly over time,” Williams commented.
What these examples also highlight is the need to embrace increasing levels of automation to build scale into personalisation programs. Because personalising 500 million digital interactions in the past year at M&S didn’t come without automation.
“You can’t make millions let alone billions of decisions on what to show customers when they interact with you without introducing huge levels of automation and taking out lot of manual processes and campaign mentality we often see in many of our retail organisations,” Williams said. “This is a journey for retailers, as you typically start from a campaign focused mindset. It will continue to play a role, but it’s these highly targeted growth journeys that are generating the value and efficiency.”
In its first two years, M&S has grown share of automation in email to 11 per cent of the total; seen the percentage of visits with greater than three personalisations rise to 25 per cent; and lifted personalised interactions from 20m to 500m.
6. Iterate and experiment
As is also clear from these examples, experimentation and constant optimisation have been essential to building incremental return from personalisation efforts.
At M&S, outcomes-focused OKRs are used to set annual and quarterly goals, supported by mission-based teams. These teams define a problem, pull cross-functional individuals together, define the OKRs then have accountability and ownership to find the solution.
“We had a mission team for product recommendations for example. We gave them a revenue and scale target, their role to work out how to get there using out of the box, building internal capability and we held them accountable to that outcome,” Williams said.
“Building an experimentation culture achieved two things. First is becoming more data-driven in deciding what we do. Too often, we heard we do it because it’s right or because a competitor is doing it, but we had no idea if it was generating a positive or negative experience.
“Also, we’re tracking velocity in experimentation as we see that a great proxy for learning and we will do what we can to unblock anything slowing that velocity.”
Victoria’s Secret has a similar commitment to constant experimentation and iteration. “Making this large shift from one size fits all marketing and digital strategy to multi-faceted, AI-driven algorithm-based approach has its challenges. The primary one was evolving processes, our people and tools and driving adoption of new capabilities,” Brunner said.
“The major key to success has been a robust test-and-learn agenda and relying on data to determine the next experiences we bring to our customers.”
An example she pointed to are individualised offers, where Victoria’s Secret has spent time unlocking how to better target offers at the right time and place.
“It’s about connecting offer experiences for marketing drivers from campaign back to site and app; targeting high-value customers using Audience Manager with a timed promotion to drive urgency to convert; and sometimes upselling customers in cart to drive higher order value if not in high value customer group,” Brunner said.
With personalised content, Victoria’s Secret constantly tests homepage plays targeting different audiences with different content. This could be acquisition messaging to non-credit card holders to sign up to be a branded credit card holder; or geo-based content.
“We continue to iterate, test and learn on all these features to squeeze out as much juice as we can and continue to optimise each activity,” Brunner added.
You can also follow CMO on Twitter: @CMOAustralia, take part in the CMO conversation on LinkedIn: CMO ANZ, follow our regular updates via CMO Australia's Linkedin company page
In the third and final episode of our 3-part CMO50 video series exploring modern marketing and why it’s become a matter of trust, we’re delighted to be joined by Telstra’s former CMO and now digital services and sales executive, Jeremy Nicholas, and Adobe VP Marketing Asia-Pacific and Japan, Duncan Egan.
Flash back to the classic film, Willy Wonka and the Chocolate Factory. Television-obsessed Mike insists on becoming the first person to be ‘sent by Wonkavision’, dematerialising on one end, pixel by pixel, and materialising in another space. His cinematic dreams are realised thanks to rash decisions as he is shrunken down to fit the digital universe, followed by a trip to the taffy puller to return to normal size.
Why is it there is no shortage of leadership development materials, yet outstanding leadership is so rare? Despite having access to so many leadership principles, tools, systems and processes, why is it so hard to develop and improve as a leader?
As a nation united by sport, brands are beginning to learn money alone won’t talk without aligned values and action. If recent events with major leagues and their players have shown us anything, it’s the next generation of athletes are standing by what they believe in – and they won’t let their values be superseded by money.
