In association with
MW: Pulling all your consumer data together sounds simple, but can it be a lengthy process?
DP: Two years ago, we started on a five-year change programme to create a single customer view (SCV), although as a charity, we see it as a single supporter view. The ultimate aim is to get closer to those supporters and deepen relationships. We’re taking a phased approach and it is taking longer than we thought, but by the start of July, we should have a single marketing view and from there, we are looking at services and beyond.
RF: We have a total customer view held within our customer relationship management system, which we are using to build a marketing system. It has been an interesting time trying to manage what has been expected of the single marketing view, looking at what was anticipated and promised versus the reality of what we are able to do.
GS: We are taking a phased approach, where we are pulling together all the marketing databases to move towards a true single customer view where we can link in with customer services and all other operational areas.
MM: We have been building a single customer view for the past six months and anticipate it will take a further six to 12 months to complete. But we recognise that there are stages along the way where we can input what we have already.
MW: So building a single customer view is not just one activity but a journey. How does that process unfold?
RF: I spent a great deal of time during the first three to six months of the project saying that we were at the stage of trying to reset the foundations of Oxfam’s customer data resources. But once you begin the process, the expectation is that you’ll have a report delivered in two weeks’ time.
We sometimes have to say that we don’t have the data or we’re not used to the system yet. Or maybe we’re inundated by having to produce 100 campaigns in the next two months. We’ve built something, but we still need time to consolidate, understand what is being delivered, how it works and what it is that people want from it.
MW: Is there still a perception within some companies that a single customer view is a solution to a problem that may not yet exist?
RF: People did tell me during our build process that they didn’t really know what they wanted from it. So you never really know what you’re trying to build. You’ve got to aim for flexibility and transparency and – having built strong foundations – you can hopefully factor in whatever you need on top of that.
MM: At BA, we held a requirement workshop around the company looking to see what data could drive in different areas of the business. On the sales and marketing side, we have clear ideas of how we could use this. It is harder to do with other touchpoints involving customers because more investment in the infrastructure is needed.
Once you start looking at the journey process for the customer, for example, you can see where the business benefit might be. But with call centres, airports or indeed on-board, using this data becomes more complex. Irrespective of which area of the business this data is coming from, though, I think it’s time for it to deliver tangible benefits.
MW: But how much information about your customers is too much?
DP: We have to bear in mind the wants and needs of our supporters as would any company. People’s motivation for giving to charity varies but essentially we are asking them to give us money for nothing in return.
About 12 months ago, when we were asking what data we wanted to hold about our supporters, the knee-jerk reaction would have been to want everything. But we actually needed to work out what was relevant for us and what our supporters are comfortable with us knowing. If we gather information that our supporters are not expecting us to use, it could harm our relationship with them.
Asking for certain data can provide strong, emotive feedback. People are not comfortable with us having certain types of information about them or their partners, for example around legacy giving. One of the main focuses and benefits of this programme to change to the single supporter view has been us talking about extending the lifetime value of the supporter. Anything that has a negative impact on that would be significant.
GS: It’s all about trying to demonstrate the return on investment for the business. I sometimes feel that when you’re specifying a data system, it’s very easy to say that you want everything included, when in reality you use 20-30% of what you’ve specified. There are so many benefits to be had from pooling data together that it doesn’t have to include every single thing you’ve got within the company.
MW: Does that mean we don’t really need a single customer view? Would a single marketing view be enough?
SG: Our business is very marketing-led. So we felt any data relating to that was a core thing that we should really have control over ourselves.
That was the driver for creating a single view in the first place.
When we went live with our aggregated data, we were able to start campaign targeting but it also gave us the ability to start modelling customers’ propensity to lapse, for example. That opened the doors for us to do other things. It meant we could move our marketing database to a customer database because we want the company to see this as a solution for the whole business, not just marketing.
But again, what we haven’t established at this stage is ultimately what we want to get from the single customer view. We want advisers to see all the interactions going on with our customers but as yet, we’re not sure how frequent we want that to be.
SA: In Q3 last year, we embarked on a project to pull together all our sources of customer and prospect data into a single marketing view.
We could talk about a single customer view but it’s more a defined prospect view. We sell our own products but also sell on an affinity basis via trade unions. Then we use other people’s products, which we sell under our umbrella. We need to make sure the data from our own business, trade unions and direct customers all marries up. This is the trickiest element. The sophistication of de-duplication techniques is one of my particular pain points at the moment.
MW: Is it enough to rely on your own data gathered internally or must you always look to third-party sources for the full picture?
SG: We use many of the common third-party sources but our issue is that there can be a six-week lead time between requesting the data and being able to use it. In the insurance industry, this can be problematic because customer circumstances may well change in the interim.
The issue of de-duplication can be solved by creating marketing rules such as never mailing more than one household, rather than a person. And elements such as customer record suppression are particularly important to get right. When moving to a single customer view, where you’ve got all the interactions together and you want other business areas to see things such as the finance department and call centre operations, seeing the same customer appear twice is not very good. We have to work hard bringing all that data together with strict match and merge rules.
DB: By accruing such large volumes of data over a long period of time, you can build up clues when dealing with duplicates that indicate you are seeing the same person’s records. It takes a combination of string-based logic, algorithms and a lot of data from various sources to be able to achieve it.
GS: The analyst Gartner claims that data decays at 33% a year. It’s really important that you use third-party data not only when you’re building your single customer view to create the right rules as to how to merge data, but also to go back on an annual basis and look at the data you’ve got. Most CRM systems fail because people are just not using accurate data or they’re not maintaining it.
SK: The ideal is that if you can use your existing customer base, then fantastic. We can do propensity modelling with our own data but in certain instances – such as realising which customers live in flats and would benefit from a core Sky installation that prevents the unsightly crop of dishes from balconies, for instance – it isn’t rich enough. To solve that problem about the flats, we looked not just at our own and third-party data, but information from ordnance survey maps too. Finding close geographical proximity between customers has allowed us to create our propensity model for this system.
The main issue is that there are so many nuances to data that you really need to tailor it to a specific business requirement. And it’s not just about the data but how clever you can be with it too.
MW: So third-party data is important, but does it matter where it comes from? How important is third-party data quality?
DP: You sometimes find there is a conflict between what a supporter has told us and what third-party data from another source tells us. We don’t currently have any business rules in place that allow us to do something about that.
Our default position is to be simple in our approach and go with the latest information we have, wherever it comes from. Our legacy data is perhaps the most sensitive so we err on the side of caution and go with whatever our supporter has told us directly. We have a lot of supporters that give via standing order or on one-off donations, so we don’t have an ongoing interaction through which to gather more data from them, which presents us with a challenge. And understanding someone’s motivation for giving is a massive piece of information that we want to obtain.
SK: When you’re looking at large volumes such as electoral rolls, you can be looking at a few million records. We rely on agencies sending data that is as accurate as possible.
RF: Oxfam doesn’t use much in the way of third-party data. We have a large database of passionate supporters and we look to that to understand their life stage and affluence levels. A lot of the models we build are around when a donation has been received, its frequency and its value.
In terms of data enhancement we have a group of people called “major donors”: generally donating more than £10,000 a year. To enable more of a one-to-one relationship with these donors, we enhance the data to try and narrow down the list. We don’t want to be writing twice to someone who is already passionate about Oxfam because they would be disgruntled by the interruption and more confused by the waste of resources when their money could be put to better use.
GS: The area where third-party data works particularly well is in product development. It’s very often the case that the product people will have something and it’s up to the marketing department to find the right people to sell it to. By having a lot of third-party data and really understanding who your customers are, you almost turn that on its head by identifying areas of your database that could be open to different types of products.
MW: Should the single customer view sit within marketing or is it vital for other departments too?
MM: As we get a richer view of our customer data, there are some strategic initiatives that require a real insight across our whole customer base. We’re using that to drive product development and customer service. It’s a great way of exploiting and building the programme overall by showing business value.
SP: For American Express, we modelled the concept of a paperless product that the brand thought would be good, but didn’t know why. We looked at how we could pilot it and use the data to build the business model. We knew it would be a good thing in terms of costing the business less and helping the environment but we didn’t know what customers would think. Only when we had the insight from the data did we know that we should go ahead.
SG: We’re fortunate that our marketing department sits within planning and strategy too, so our marketers are able to share the information that we’ve got. I think when the business signs off the costs of a project such as single customer view, it is anxious to see some results. Marketing campaigns are an obvious place for this and so they seem to come first. But thereafter you can start to engage other areas of the business.
RF: The single customer view tends to sit within the marketing silo because the first thing you want out of it is an increase in response rate or decrease in costs. The single customer view allows us to see the supporter journey, which is particularly important if they start their relationship with us on a non-financial footing, such as signing petitions or taking part in marches. These supporters may not be bringing money in now as student activists, but we know from analysis of our customer databases that they may well do so in the future and it helps sell in the data imperative as a business case.
DP: Trying to define non-financial value is certainly an issue for us and the expectation is still very much on our marketing function. When wedecided to invest money in this a couple of years ago, we knew it was a lot of supporters’ money and you have to justify the benefits over the long term.
Immediate benefits are selection, cost savings and sophisticated targeting. That doesn’t feel like the big prizes but they are the most immediate ones. Now we can see the supporter landscape through this, it’s helping us to be more strategic about longer term value. We want to show the rest of the business this is a cultural thing and everyone needs to realise its importance.
MW: So is the business case for a single customer view justified by the gains from better marketing or the long-term benefit?
GS: It’s a mixture of the two. If you’re trying to sell anything to the board, it’s got to be based on how you are going to get a return on investment because it’s a big step for any organisation. You start involving other areas of the business such as IT when companies realise how much they have underestimated the cost of holding data in a variety of different silos and formats.
When trying to justify putting a new database in incrementally, it may not be seen as the best solution because you’re already paying a number
MW: Is it the marketing department’s responsibility to carry the cost of implementing this programme?
SA: It is a direct cost of our marketing activity, although in recent times I’ve been impressed by the number of suppliers willing to work flexibly to allow cost and its associated risk to be spread over time. Some are perhaps delivering a lower upfront cost with a slightly inflated monthly fee as opposed to the need to pay out hundreds of thousands of pounds upfront. This way, the value of the project is built up over time by both parties involved. It’s refreshing to see how things have developed and people are willing to work with you on a project.
SP: There are still the issues of deliverability and timescales. Some people ask the board for money and claim they’ll see a return in three or six months, but in reality it’s very difficult to see any form of return in less than six months.
And though you have a new piece of technology, you don’t start changing the way you market instantly because you’ve got to re-educate your people and start thinking differently. What you find is that people undergo training but you’re still doing the same sort of thing. Realistically, you can be looking at a nine to 12-month period to really start seeing those benefits.
MW: The Data Protection Act increased fines for non-compliance from April 2010. Does having all your data in one place give you cause for concern now the margin for error is so low?
SK: The risk has always been there – I think the population as a whole is already well aware of its rights in relation to personal information. What we have through the single customer view is fundamentally just another source of data.
We’ve just got the one repository that marketing is able to use better and we’ve got to make sure that we’ve got the governance in place to protect it. It is always important to have support and steer – particularly from the legal arena – throughout your whole business.
RF: We take data security very seriously and we have a supporter relations team internally that deals directly with the customer. Because not all of our supporters are financial ones, perhaps the checks aren’t always there. Also, we have a fairly open-plan environment, so we have to be careful what is left on desks or on screens, but we conduct an annual review of data security and are constantly looking to improve.
MM: In terms of data security, I don’t believe the single customer view provides any more or less risk than any other form. The same stringent security measures are required.
GS: In some respects, by looking to develop a single customer view, the most complex part is integration. When you are at that stage, it is like a great data audit of what your company stores. It does raise the question for all the individual sources – do we need to hold as much customer information as we have? Good brands should always review the information that they are holding and methods they are using.