Highly targeted and personalised ads may not translate to higher profits, according to research published earlier this year by professor Pedro Gardete at Stanford Graduate School and Yakov Bart, a professor at Northeastern University, Massachusetts. In some cases, they claim, the most effective strategy is for consumers to keep information private and for businesses to track less of it – which is the opposite of current thinking.
The pair used game theory to build a mathematical model enabling them to test the boundaries of “cheap talk” – for example an ad promising the “lowest prices in town”. They found that the most personalised ads were “less effective because consumers worried they were being exploited”. They were also left frustrated. Gardete cites the example of someone looking for a prom dress and receiving an ad claiming ‘We have a wide selection of prom dresses! Click on this link!’ But when they do, it turns out the retailer has dresses for all occasions but not specifically proms.
It might therefore be better for brands to collect less customer data, he explains. “It’s to companies’ benefit to know less about consumers and also disclose up front on their site exactly what they do collect because consumers don’t want to feel a business has tailored its ads to them based on what it knows about them. That feels manipulative.”
It’s an interesting theory, but does it stack up in the real world? Do the majority of advertisers really run the kind of campaigns Gardete and Yakov analysed and, if so, would the collection of less data result in more customers?
Only collect what you can use
“If you look at the rise of ad blocking software it’s rather obvious that there are ads out there that people don’t find relevant,” says Annabel Kilner, commercial director at furniture retailer Made.com. However, the orchestration of data, planning, and technology maturity required to ensure highly targeted and meaningful messaging to a “segment of one” varies from one situation to another – and depends on ad format and where someone sees the ad.
Online display advertising, for example, is “lagging behind”, Kilner continues, but search marketing and social advertising are becoming “increasingly relevant” to consumers as additional data layers become available to improve message relevance. “Advertisers that have a credible brand; that have compelling and honest messaging in their ads, which is refined based on resulting sales; and that effectively use site data to continuously better understand what customers intend to buy will experience higher ROI from their personalisation and retargeting efforts,” Kilner explains.
Lindsay Grieve, head of customer management at Macmillan Cancer Support, puts it this way: “The old way of shouting very loudly above the line and hoping it sticks still has a place but we have to get better at talking to people and understanding what they want. You need the systems to do it; you need to capture data. But don’t collect data you’re not going to use, which is what I sense happened in the Stanford [research report] examples.”
Zak Hood, head of ecommerce at outdoor clothing and ski specialist retailer Ellis Brigham, fundamentally disagrees with the main finding of the study – that in determining the tone of their targeted messages advertisers “may prefer to remain in a state of partial wilful ignorance [about consumers’ preferences] so as to preserve communication credibility”. One of the key advantages online marketing has over offline is that everything you do is quantifiable, he explains. “Why would anyone want to remain wilfully ignorant of whether or not their money is being spent wisely?”
Empower consumers to find the right product
Ensuring the marketing spend is put to work efficiently is one thing. Where data really creates a step-change in relevance for customers is when it’s used to ensure highly personalised ads that lead to a sale. Then data is then also used in the post-purchase experience, and that can turn a customer into a repeat customer.
Research showing that consumers want a personalised shopping experience is plentiful. Hardly surprising, then, that the top digital tasks for executives in 2016, according to Adobe’s February 2016 survey with Econsultancy, are increasing targeting and personalisation. Of the 7,000 global marketing, digital and ecommerce professionals they quizzed more than one-fifth (22%) cite experience as their priority. Data is the oil that will grease the wheels, putting marketers on the road to achieving both.
Targeting is more than using a hammer to “whack the alien”, as Franz Aman puts it, referring to the frenetic amusement arcade game. Aman, co-author of the book The Marketing Data Lake, says: “It’s important to understand the customer fully and to be helpful.” He adds: “Marketers need to empower the consumer to find the right product, not trick them into buying something else.”
Retailers, for example, are using big data to forecast demand during the busiest times of year so customers aren’t left without the present they ordered at Christmas or the bargain they’d banked on during Black Friday.
German fashion retailer Otto has been posting some impressive results on the back of its investment in predictive data analytics with Blue Yonder. Using around 200 different input variables – including brand, price, online placement, stock situation and weather – the team can predict the sales probability of items every week for the whole season with demand updated on a daily basis. This equates to five billion individual forecasts every year, but forecast quality has improved by “up to 40%”, says the Otto’s vice-president of category support, Michael Sinn.
These precise results help the company plan and manage its selection of products for the future, as well as reduce out of stock rates and increase revenue through replenishment and price optimisation. That’s worth its weight in gold in a sector characterised by low margins, high competitive pressure and fast-changing market conditions and customer demands, explains Sinn. “You’ve got to remember that customers are only one click away from the next competitor.”
Evander, which supplies and fits a range of doors and windows, is also using data to help forecast supply and demand – especially during busy periods. One of the company’s promises is to carry out a building survey within 48 hours. “We use historical campaign data to ensure we have the right surveyors in the right areas of the country to deliver that service,” explains the firm’s SEO and marketing leader Robbie Fowle. “For example, over 60% of our business is based in the South East, so in key months we will move resource from Scotland to London.”
‘Less is more’ in retargeting
The use of geo-location technology to track customer movements is attracting a lot of interest. The contextualised marketing that comes from this can border on the creepy, so brands certainly have to ensure they aren’t making people feel uncomfortable. But there’s plenty of research showing that shoppers would appreciate offers being sent to their mobile phones whilst they are browsing around a store. Trying to keep up with where a potential customer is on their purchase journey isn’t easy, though.
Retargeting campaigns target customers once they’ve left a brand’s website and are now a standard part of the online marketing mix, but they often become intrusive and a ‘less is more’ approach may be relevant, says Ellis Brigham’s Hood, who works with personalisation specialist Persomi.
Researching a product on a phone and then buying it later on a laptop is common these days, for example, but because most advertising platforms have no fool-proof way of tracking this journey across devices “you could end up following the visitor around the internet, delivering ads to their phone for a product they already purchased several weeks ago”, he explains. “This can seem intrusive [because] the message is no longer relevant and the result can damage your brand, particularly if the visitor never asked to be marketed to.”
Sarah Boustouller, head of marketing at Stephensons Solicitors, believes ‘less is more’ is a relevant motto given the rise in the use of ad blockers. “For many this does mean taking a step back as brands once again start to focus on raising general awareness, rather than data-driven, targeted campaigns promoting specific services or products,” she says.
Rejecting targeted communications in favour of mass marketing might not always be the best approach, noted Melbourne Business School associate professor Mark Ritson in his column for Marketing Week earlier this year, written in response to comments by Mars CMO Bruce McColl arguing the opposite.
“There are still many instances where a clearly identified target segment will make you more money than a mass marketing approach,” he wrote. “I still teach targeting as an explicit strategic choice. Do we target everyone like Mars? Do we target a couple of segments? Or do we make the leap of faith that says because of our size or the market’s dynamics we will only go after one segment?”
Collecting data that enriches brands’ understanding of their customers’ intentions and needs can serve to improve visitors’ experience on-site as well as improve the relevancy of communication they receive off-site, says Made.com’s Kilner. “Without data about users, all experiences would be homogeneous, and that can’t be a good thing.”
The deadline for this year’s Data Storytelling Awards has been extended to 2 August. Click here for more information on how to enter.
Expert insight: How Made.com and Macmillan Cancer Support are using data to better target marketing
Annabel Kilner, commercial director, Made.com
We use data for prospecting and messaging to people who we think would value our brand and products, as well as retargeting for a period of time to people who have expressed an interest in our brand or products.
For prospecting, we use data layers from our partners to target lookalike audiences we would like to message, who have not experienced our products yet. For retargeting, we utilise prior interactions – for example a product view, add-to-basket, or purchase of another product type – to message people with highly relevant ads tailored to them.
The targeting takes into account many data points including how recent the customer’s interaction was, as well as how often we’re showing them the ad, to ensure ads are relevant.
Lindsay Grieve, head of customer management, Macmillan Cancer Support
The old way of shouting very loudly above-the-line in the hope that it sticks still has its place but we have to get better at talking to people and understanding what they want.
I am horrified [when I discover that] we have emailed some people three or four times with different propositions. I registered for everything [we offer] and got six messages, which suggests [Macmillan] doesn’t know what I want. Equally, if you send someone something three times and there’s no response then it’s clearly not resonating with them – the data can tell you that.
We’re getting better – we are becoming more targeted and more precise. For example, a coffee morning won’t appeal to our younger supporters, so we’ve been able to [target that segment] with new events like Tough Mudder. I’d give us a seven out of 10 currently, but it’s a journey.