If there is a cure for problems in a business, it might well be found in the use of data and analytics. That came across clearly in the 2014 DataIQ NOW! Conference, where examples of business transformation through intelligence were freely dispensed. By David Reed.
What would you do if your segmentation was generating a quarter of a billion customer types? If you had selected a CRM vendor and then discovered during a test of the system that your business could not deploy what it was telling you, would you have the courage to start again? If 80 per cent of your target market visit your stores, but 70 per cent also visit via mobile phones, how would that affect your customer experience management?
These were some of the challenges faced by speakers at DataIQ NOW! Who offered learnings from their real-life experiences to a buzzing crowd of over 260 delegates. Whatever the hype might be around the latest data types or technologies, these practitioners provided authentic knowledge gained from “boots on the ground” experience - and how data and analytics may not be a cure-all, but they are certainly able to improve the health of a brand.
Opening the event, Mike de Halpert, head of EU CRM and loyalty analytics, eBay, noted that big data is attracting investment, but not always intelligence. “You can throw money at the problem of big data. But the real difficulty is with complexity and how that affects your business model,” he said.
As an example of this, he recalled his own reaction to the data assets he discovered on moving to the online auction site. “When I joined eBay, I was excited about the amount of data and talked to the marketing director about doing customer segmentations. I expected him to get excited and see it as a quantum leap. Instead, his reaction was that, if we created just five segments per attribute, we’d have 224,140,625 customer types - the business can’t handle that,” said de Halpert.
That level of complexity has to be challenged, even though, as he pointed out, “there is no science of simplicity”. De Halpert offered five rules to help: the first of these was to value simplicity because, “if you don’t, things will run away from you.” To help avoid that, his second rule is to align big data goals with company goals. “How often do your analysts go off on an individual frolic and discover correlations that are useless to the business?” he asked.
Rule three was to draw up a roadmap to success that shows how the business will progressively adopt analytics, while rule four is to standardise everything. “Before I joined, everything was bespoke and analysts were doing things to please the marketing department - they were being too helpful. So the first thing I did was to standardise. Marketing didn’t like it, but now they can run 1,000 tests per month, each generating five or six pieces of information,” he said. His fifth lesson was using visualisation tools to ensure insight is comprehensible.
Many of these lessons were reflected in the presentation given by Paul Laughlin, a 24-year veteran of analytics and former head of customer insight at Lloyds Banking Group. He revealed how the bank had created an operating model which provides the business with the insight it needs and also improved its business intelligence. “In the past, modelling had been constrained by not having good data on what the organisation did, not just on what the customer did,” he pointed out.
This shift has been critical because of new financial services regulations that require banks to show an offer was based on customer needs, not profit targets. “Do we have evidence that customer behaviour has shown there was a need and that they fulfilled it with us rather than the competition?” he asked.
One outcome of this shift which is not typically found in analytics has been “a targeting strategy that allows for the customer to be left alone,” says Laughlin. Equally, it has seen efforts made to fill in a blind spot where the customer expressed a need, but LBG was unaware of it. “Before we go out to somebody with an offer, we now apply a filter - is it appropriate and do we have permission? Then we look at the uplift in outcomes that delivers,” he said.
Selling this change in thinking to the business has not been easy, not least because short-term sales targets can often disrupt what analytics says would be best. But the uplift in terms of accuracy and quality from the insight programme has been in the order of a multiple of 2.75 overall. “It is early in the journey and it is not complete, but the results are positive,” said Laughlin.
Changing views within a business or marketing function is never easy, but having evidence from data or analytics can often help. At last year’s event, Comic Relief showed how insight from DataTalk had identified the impact which door drops could have when delivered to neighbours of an existing fundraiser, increasing the average sum raised by up to 600 per cent. This had persuaded its marketing director to use the channel, despite wanting to move the activity fully digital.
For 2014, the charity combined supporter data with school and retailer data but with the caveat, sounded by DataTalk’s managing director Tim Drye, that “analysts need to remember that underneath the numbers are people. No database is ever complete.”
With this in mind, Comic Relief focused on social behaviour. “When somebody does something on your street, others are more likely to follow. People are not just individuals, we live in herds and 50 - 60 per cent of people are not risk-takers, they prefer to follow,” said Drye. The charity reflected this in its fundraising material, which included window posters and cards which supporters could hand out to others.
“The door drop emphasised the local nature of the campaign,” said CRM manager Liz Curry. The result was an indexed uplift of 150, with hotspots of 380 in areas with five or more people fundraising in the same street. Overall, Comic Relief raised a record sum of more than £51 million in 2014.
Change was also a theme for Rachel Hall, head of CRM and database manager at Honda (UK). She explained how the company had sought to address a weak CRM programme and lack of insight by adopting a new solution, only to discover that it might not solve the problems. “We followed the Honda-approved vendor selection process,” she noted, with a cross-functional business team that identified a solution.
“I requested training for some key staff ahead of purchase to ensure the tool was as intuitive as it appeared. That highlighted it was easy to use and would identify more targetable prospects, but we realised there were serious data latency issues. Users had no training in statistical analysis to make sense of the output,” said Hall.
Honda’s existing customer database did not sit in the marketing function at that time which amplified the potential risks. “Our vendor review had been a technical process without business buy-in - we thought we could get that retrospectively,” she admitted. A further proof of concept from the vendor revealed a new set of issues with both the business and its data - the database function had come to be viewed as “the department of ‘no’.”
So Hall started again, fortuitously just as a new chief marketing officer joined who bought into the value of data and ensured the database got moved into her function. By improving the data and its outputs, combined with a strong internal communications effort, the situation was recovered. “The function of the database is to sell more - the people working on it need to focus on that, which hadn’t been the case,” said Hall. Within the new process and environment, data now flows more rapidly towards its users, with a reduced cycle time from two months down to two days.
Another business transforming itself in a changed market is Mothercare. The brand has been through challenging times with profit warnings and the departure of its CEO, yet has managed to retain a place in the affections of its target audience. “Eighty per cent of first-time mums-to-be go to a Mothercare branch,” head of CRM Ryan Davies told the conference. “Trust, time and location are key to the customer journey,” he said, especially as a result of fragmentation of channels with most parents using their mobile phones to access its website.
The focus of creating the right experience has been to ensure data capture of the child’s data of birth and the creation of a mobile app and email programme that is actually useful. “Our insight was that the needs state of parents-to-be changes so quickly, within six months they are in parenting mode and need different products and information put in front of them,” he said. “So we needed to go through all of those lifestages and improve the relevance of our marketing and how it was delivered across channels.”
Mothercare has the advantage of knowing exactly how a child will develop and what products are relevant in its first six years. Applying this to lifestage emails had a significant impact. “We ran a test in 2013 which increased the funnel of visits to our website from every 100 emails from five to 23,” noted Davies. By combining lifestage and event triggers with rich relevance on the website, the brand is now seeing significantly better engagement with parents.
Crucially, the business had discovered the value of friends and family who might be influenced by parents. This “communicator value” might show that a customer who spends £140 with Mothercare also brings in a further £280 from their social circle. Likewise a pure influencer who only spends £50 directly could be responsible for generating up to £1,730 via baby showers, birthday gifting and the like from within this circle. “We are creating a connected experience in a traditional business. Customers don’t see our silos, they only see the brand,” said Davies.
As these experiences show, businesses are steadily being transformed by their adoption of data and analytics. But this is not just a case of managers moving towards this industry - it is also driven by data practitioners making themselves more business-friendly. As de Halpert explained, every one of his analysts is partnered with a business manager to ensure they retain a commercial focus.
That point was further emphasised by Gary Childs, head of insight at Callcredit, headline sponsors of DataIQ NOW!. “Five years ago, analytics was seen as black art - nobody knew how to use and deploy it. We are now seeing it move from being prescriptive to predictive,” he said.