David Lockwood from The Tapestry Agency explains how to analyse your customer data

David has an impressive career in marketing and analytics, working with companies such as Boden and Direct Wines before co-founding The Tapestry Agency.

He speaks to Andrew about customer-centric marketing; how it’s normal to lose money when acquiring customers, how David analyses customer data to determine how those customers are going to perform for the business; and the importance of triggered emails when targeting customers.

Also available on:
Apple PodcastsSpotifyGoogle PodcastsAmazon Music
or from your smart speaker.

Andrew Veitch: Welcome to the Joy of Marketing with me, Andrew Veitch.

This week, I’m joined by David Lockwood. He’s had a long career, sorry David, I didn’t mean to make you sound overly old. You’ve had a long career in marketing, including a spell at Boden as Marketing Director, and he is now the co-founder of The Tapestry Agency.

Welcome to the show.

David Lockwood: Thank you very much.

AV: So David, can you start by explaining customer centric marketing?

DL: Yeah. So in quite a lot of places I’ve worked, and quite a lot of the people that we work with, people are, marketers are overly focused or extremely focused on how their campaigns work, that they set themselves up. And they set up most of their measurements around spending money on a campaign and looking at the return on that individual campaign, without necessarily thinking about, who are they selling to, and who you’re aiming these campaigns at our customers, either existing customers or potential customers, but it’s all grouped around groups of customers.

So it’s thinking around, how am I measuring the impact of all of my marketing efforts on customer groups? Rather than just measuring the output of campaigns, because multiple campaigns will roll up to have multiple impacts on customers, rather than just the individual campaign.

AV: Yeah, and that’s just probably part of a bigger thing that in marketing people generally are quite obsessed about customer acquisition. And I would actually, in fairness, I suspect that when we were with CMOS ourselves, I mean, I certainly know I probably fell, I probably made that mistake, myself, too. But I mean, in terms of that long journey of a profitable relationship with a customer, and when the acquisition pieces it’s just the start of it anyway.

DL: Yes. And I think most brands would agree that they make the majority of their profit from those customers they retain, and those customers, they have long and valuable relationships with. The recruitment, you absolutely need to obsess over because if you’re not bringing new customers into the brand, then you’ve got no potential of nurturing that relationship, and then turning those into long term loyal customers, ideally, even advocates for the brand. But it has to start with recruitment. And that I think, is why why we perhaps obsess there a little and you can perhaps get over obsessed about that and forget about the fact that the engine room, the powerhouse of your brand for most people is the loyal customers, and measuring what you’re doing to those customers and how you’re impacting the outcome of what you do to them.

AV: Yeah, and I suppose I mean, thinking about most direct to consumer businesses, I would say the majority probably lose money on order number one, and then it’s order number two onwards, you know, where they begin to hopefully make some profit.

DL: Yeah, I’d say, with the clients we currently have, probably two thirds of them lose money on recruiting customers into the business. Certainly the brands I worked at most recently, where you’re acquiring very big numbers of customers a year like, Boden and like Direct Wines, they’re losing reasonably substantial sums on the first, on the first order on getting that customer engaged in the business in the first place.

The really important thing is understanding how quickly that loss turns into profit. So having a really good idea on how quickly those customers start to return to you, that the investment that you’ve made into them. And that’s all around, again, customer centric analytics, understanding how long it takes from a new customer coming into the door to them placing a second, third, fourth order, and what those order values tend to be. Do they tend to climb? Do they tend to shrink? That, you know, what are all those metrics? And how do they shape up into turning a potentially loss making new customer into a profitable existing one?

AV: Yeah, suppose that’s one of the difficulties is, in a way, you can only tell whether a campaign worked quite far into the future. And I suppose expanding that out to the whole business, you know, a bad direct to consumer business will lose money on the first order, and then never really make it back; a good one will lose money on the first order and make money in the future. And I suppose in both of these situations, it can be quite hard to tell whether the campaign or the business, as a whole, is going down a good path or a bad path until you know, quite far into the future.

DL: Well, I’m not I’m not entirely sure about that.

So the majority of the businesses we work with, their cohorts of customers are incredibly predictable. So any given cohort of customers we can gauge what their spend over a 2/3/4 year period will be within the first 60 days of a relationship with that new customer cohort into the business. So I don’t think you necessarily have to have, you don’t have to be looking that far in the rearview mirror to really understand how strong that campaign was.

The other thing is, while you do have variances in the future value of customers, depending on how you recruited them, and what media you used, and what product they came in on, and all those kinds of things, actually, the future value of the customer tends to vary less, significantly less than the cost per recruit of getting a new customer into the business does. So you know while the cost of recruit in a new customer into the business might vary by 80/90% plus from one campaign to the next from one media to the next, actually, the future value of a customer is likely to change by 20/30/40%. So actually, it’s kind of much more important to get the understanding upfront of what your costs per recruit is, you’re definitely going to need to understand the future value of those recruits. But any given cohort of customers is likely to vary significantly less than the in the upfront cost will vary.

AV: Sure. And just to step back, so you said there within 60 days, you can generally have a pretty good view of how that customer is going to perform.

So what are the factors that you’re looking at?

DL: I look at it in a couple of different ways.

So what percentage of those customers have returned after 60 days? And what percentage of customers, using historic data, what percentage of your customers who return at 365 days, 720 days, 1080 days, 1190 days, whatever number it is, you’re looking at. What percentage of those customers have returned at 60 days?

So if as a business normally 30% of anyone who’s ever going to repeat purchase from you has done so within 60 days, you can work out whether that’s a figure which is normal, you know, does that always stand true within the business? And then you can then calculate that out.

So it’s looking at how many customers return and then it’s looking at how much in total have that cohort of customers gone on to spend. So again, you look after 60 days, how much is that cohort spent after 60 days using historic data? How much, how much have cohorts historically spent after a year, two year, three years, four years? And working out what percentage of the 60 day spend is of the future spend and then looking, do you have a trend that is predictive? And almost all businesses I’ve ever worked with have a predictive trend after. The longest I’ve ever found it was 120 days, the shortest I’ve ever found it was 15 days. But most businesses should be able to predict the future value of any given cohort of customers relatively quickly.

AV: Yeah, and I suppose, I’m guessing that that’s particularly the case in businesses where it’s more of a sort of consumable product? And I suppose food and drink is something you’re going to probably spot quite quickly, whereas maybe fashion perhaps a little bit a little bit longer?

DL: Again, I think there are a few things you understand you need to understand first.

So what’s your time to second order for your business? So if you recruit 100 customers, you expect and your business expects 50 of them to place a repeat order at some point in the future, at what pace do they come back? What percentage of those people have come back in 1 day, 5 days, 30 days, 90 days, 480 days, you know. Whatever that curve, and understand what that curve is, and what you’ll find, again, for most brands, and I would exclude perhaps some of the furniture brands in this, but all of the fashion brands I’ve worked with the likelihood of return and the speed of return, so your golden window of opportunity for bringing customers back, is actually much quicker and much shorter than most brands perhaps would expect before running the analysis.

AV: Yeah, I suppose something I often worry about too, is when customers , I mean clearly it’s correct that you want to take a view on lifetime value, and then maybe invest more in recruitment, if you’re seeing good lifetime values. But I do sometimes get concerned when people just go a little bit too far down that route and the gold is maybe just a little bit too far at the end of the rainbow.

I mean, do you wait, you know, six months for the customer to become profitable? You know, one year, two years? I mean, just just how far can you actually really go and still be comfortable with these longer term predictions?

DL: It really depends on how predictable your future revenue from those customers is. So can you predict within half of 1%, after a very short term, how much you expect every single one of your customers coming into the business to spend over time?

I worked, as I said, at Direct Wines and their want and desire to get new customers into the business and they were prepared to invest over a much longer period than most other customer, most other brands I’ve worked with. So they would, they would invest and not expect to make profit at the time I was working there for up to three years. Now, that was a continuity business and continuity businesses, as we know have a much more reliable and much more predictable future stream of income and so your likelihood of being kicked sideways by something and not achieving your three year revenue targets were very low.

Most of the businesses I work with are looking to invest at somewhere between 6 and 18 months. Six months is a one that’s commonly used, because if your recruits are recruited evenly across the year, and you’re investing six months, it means that the ones you recruited at the very beginning, became profitable and are making profit after 12 months; the ones you recruit in the middle are breakeven by the end of the financial year; and the ones that you recruit at the end have lost money. So hopefully over the full year, all of your recruits about breakeven. And that’s why a lot of people use six months as a kind of finger in the air to say, therefore, my entire cohort of recruits in any given financial year will break even within that time period.

AV: Great. So we’ve spoken quite a bit about customer data in terms of forecasting lifetime values, are there other uses of customer data to help people understand what’s going on in their business?

DL: Yes, a lot.

We do quite a lot of work on next best action, which is understanding how customers come into the business. So how new recruits come into your business, what products they buy, or what product categories they buy and because of that, what product category are they most likely to buy into next. So it allows you to dial in some relevance and personalization in your communications, right from your third, fourth or fifth communication to those customers so that you can really start driving them down a journey that should deliver better results, because you’re understanding what they’re most likely to do next. So that’s certainly one.

Understanding what type of customer goes on to become your best customer and therefore allowing you to change the, or pivot the axis of your recruitment to make sure you’re focusing your recruitment more on those customers who are most likely to drive the, you know, most likely to drive your hero customers in future, the cream of the cream customers to come.

You can use, there are so many uses it almost becomes a little bit difficult when you when you start listing them out. But understanding a customer’s channel preferences, so that you’re talking to them in a channel that’s most likely to generate future responses from them.

All of those things come can come out of the data or can be given as clues in the data to allow you to understand what testing you might do in future to really drive home those, that data advantage of knowing what’s most likely to happen next.

AV: Yeah, I mean, stepping back, I know someone who actually runs a small wine shop. And of course, you know, he’s a huge enthusiast, he speaks to all his customers, he actually knows his individual customers’ preferences, he knows what to recommend to people. And then we know obviously, when he’s coming to select as range, he actually knows, intuitively knows, what customers are likely to buy, because he’s so close to them. Whereas I guess, you know, the difficulty we have in direct to consumer, is we don’t quite have that same close relationship. So I guess really what a lot of, what’s your and I guess what I’m trying to do my business too, is use data to help the DTC brand get into the position of you know, my friend who runs a shop and speaks to his customers every day.

DL: Yes, I mean, it.It gets very difficult when you’ve got 100,000 new customers coming in through the door every year to obviously know those customers on a be able to shake their hand and know them face to face basis. However, you can let the data give you as many clues as are in the data. And those clues can be from as basic as gender and geo location.

So, do people in different parts of the country have different preferences, for different fashions? For example, when I was working at Boden.

Or different wine types for when I was working with Direct Wines, and they certainly did. But on top of that, how did they find you in the first place? What channel did they use? Did they use a discount code or not? Did they buy a product that was on sale, or one that was at full price? Did they buy a new product to the range or one of your stalwart products, that has been a kind of hardcore part of your range forever? What day of the week did they shop on can be really indicative of what they’re going to do in future. Did they buy one product or multiple products? And what was the average selling price of the product they bought?

These are all indicators that allow you to understand how you can drive a relevance of communication to those customers at later stages. And it’s using those in combination that should help you to drive that relevance and personalization that we know. And all of us should know, and drive a deeper and more meaningful relationship that should drive a more cash positive relationship for the retailer.

AV: Yeah, absolutely. So I guess we can’t really be talking for too long about customer data without talking about database marketing, which I know is a subject close to both of our hearts. And I guess that, probably, is I’m gonna say that is that where you spend most of your time actually doing at Tapestry, database marketing of one form or another?

DL: Yes, in the kind of broadest sense of the word.

So we do an awful lot of taking the customer data and trying to solve strategy problems, tactical problems by understanding the data better. So if you’ve got a tactical problem that your AO, average order value, has been drifting down for two months and, why is it particular customer groups? Is it items for order average selling price to different customer groups? And if it isn’t, then is it a change in product mix to different types of customers? So yes, it’s database marketing from all of those standpoints.

And then it’s, so what do I do about that? And, and database marketing allows you to contact some customers directly, to try and solve some of the challenges you have as a business or try and generate more revenue per customer you have on your database. And that happens through a number of means.

I think the one that people, the one that springs to people’s mind most readily is probably email. Most people are collecting email addresses; most people are giving themselves permission to email those customers, however they collect those permissions, whether that’s legitimate interest or actual defined opt in. And then using email to talk to customers about new product ranges, existing ranges, offers promotions, etc. We found that probably, most brands we work with, on an email basis, probably use what used to be called at Boden “spray and pray emails.” So throw it at everyone and hope someone responds emails without being quite as targeted or quite as relevant as perhaps they should be.

And one of the problems with that is, where do you start? If you’re currently sending one email to everyone how can you break that down into something that makes sense? And, and you know, I’ve only got the same amount of resource I have, I had yesterday, so how can I create 15/20 different emails to allow all these emails to be personalised and relevant to individual customers?

And I think one of the places most people forget that they are already being relevant in their communications to customers on email is triggered emails. So you’re almost certainly triggering some abandoned basket or abandoned browse, you’re almost certainly triggering some thanks for your order. And you’re almost certainly triggering thanks for signing up. And you’ve probably following that up with and if you’re not, you should be, some kind of nursery programme. So what are my first 5/6/7 communications to the customer? And what are, what is the job of those communications? Are we telling the brand story, the range story, the benefit of being part of us, why you should have chosen us over anyone else, have we told all of those things and have we put that story really robustly through our nursery programme?

So you probably are doing some of the relevant stuff already.

It’s then saying, well, for the other emails, the kind of big emails, you know, the email I send out on a Wednesday and a Friday, every week, how can I start to sub segment that? And the answer to that is, do one more email next week than you did this week. Only one. So, sub-segment your email in two groups and find the two most obvious ones. For example, if you worked in the wine industry, do you have people who only drink red and people who only drink wine? If you’re in fashion, do you have people who only buy block colours and people who only buy prints? Or do you have people who only buy dresses and people only buy trousers? Whatever it is, and and you can translate that to whatever industry you’re in and just slice it down the middle and say, do I make more money by being relevant and personalised?

So it’s that kind of first age of relevance and personalization in database marketing that allows you to say, can I, can I do that and make more money?

AV: Cool, I’m gonna say, as someone that whose day job is actually running, mainly an email marketing business, some thing I will say about triggered email, which isn’t generally well known, but actually, the deliverability is a lot higher. When you send a big batch of emails, inevitably deliverability, and again, by deliverability, I mean, the number of emails that goes into the inbox as opposed to hitting the spam folder, you will always see more emails going into the inbox from a triggered mail than you will from a campaign.

And actually that at Machine Labs, it’s even more radically different because we use a different mail server for these transactional mails than we do for the bulk marketing mails. So you might actually even get as much as 5% extra hitting the inbox simply from it being a being a triggered mail, even before you go to the extra benefits from the personalization and it being being part of a programme.

But regardless of which ESP you use, if you use one of our competitors, you still will get higher deliverability on an automation than from from a campaign.

DL: And I think people perhaps under think the amount of trigger points that they could use. So are you sending a an anniversary purchase email? Which is “you first purchased from us five years ago we’d like to celebrate that by thanking you for your customer forever.”

Are you sending, if you’re collecting date of birth, are you sending a birthday email a couple of weeks or so before their birthday saying “how about you treat yourself?”

Are you sending a an anniversary purchase of a particular type of product, “this time last year you bought sandals, is it summer again, perhaps you need..” You know? So there are many ways you can set up triggers for your biggest categories of products, on your biggest types of and your biggest anniversary points to really send out a lot more triggers. Have you got triggers for when a product goes on sale? You’re sending an email to everyone who’s looked at that product in the last six months.

You know, there are lots and lots of additional trigger points that people can use. And we know that triggered emails are, as you’ve just said, more likely to land in the inbox in the first place. And secondly, so significantly more likely to get responded to because they have an element of “oh, yeah, that’s addressed directly to me, I get that.”

AV: I think virtually every direct to consumer business I’ve spoken to is doing some form of email marketing, however, only a very small minority are sending out traditional catalogues or doing anything by post.

I mean, do you think, I mean is, do you think that direct mail is just sort of a thing for the past or just the older audiences? Or is it something that that still has a has a place in a marketing plan?

DL: I’ve come from a history of catalogues and catalogue marketing. But actually, I’m a huge devotee of digital marketing and the efficiencies digital marketing can bring.

However, we run a lot of testing with a lot of the brands we work with, to understand the incremental uplift of sending paper based marketing and the incrementally, in almost all cases, is very, very strong.

So the lift of sending some paper marketing against not sending paper marketing is huge.

It also allows you to reach some people who perhaps you can’t reach through email. So the chances of someone having opted out of an email from you is significantly higher for most businesses than the chances of you not being able to contact them by post. So it broadens the reach of customers you’re likely to hit.

It definitely isn’t just for the older customers, we’ve run testing incremental holdout testing on different customer age groups, from 18 to 25, 25, to 30, 35, in bands going upwards, and actually, we don’t see a huge change in the incrementally. The older the customer is, in fact, the latest results suggested that the youngest of our customers, on some of the brands we work with, are more likely to have a higher incremental response to the catalogue than the older customers. We’ve surmised at the moment and we don’t know for sure, but we think that’s because they probably don’t get as many items in the post as the older customers. So the items they do receive in the post probably get higher cut through. So the older customers are probably still likely to be getting bank statements and you know, yeah, utility statements through the post, the younger customers, almost certainly not, there’ll be doing all of that digitally. So stuff that arrives in the post probably gets a bit more attention than it does from some of the older audiences.

AV: I remember Johnnie Boden, the founder of Borden, saying that as long as human beings use the lavatory, they will still read catalogues.

DL: I’d argue, I probably wouldn’t go quite that far, although that does sound exactly like Johnnie’s sense of humour.

As long as they’re using public transport, as long as they’re sitting in their cars waiting for their kids to come out of school; as long as they’ve got idle time and hands spare, the chance of them wanting to pick up something physical and flick through that I think are quite high.

And we used to spend quite some time at Boden obsessing over the size of the catalogue, and would it fit in the average handbag. So, you know, it’s important to us that, it was important to us that, at the time that people had free time that they were able to use that, and use that to browse the Boden range and to really get an understanding as to what Boden looked like this season. And the same was exactly true in in wine. And the same is exactly true with a lot of the other businesses we work with at the moment.

AV: So David, thank you so much for coming on. It’s been an absolute pleasure. I’ve learned a lot, and I’m sure the listeners will have learned a lot too.

So thank you very much.

DL: Thank you for having me on. Thank you.

AV: A great way to get started on personalization, if you have your own E-Commerce Store, is to use Machine Labs.

We can recommend exactly the right products to your customers by looking at their purchase history, what other products similar customers have purchased, and 950 demographic variables, including age, wealth, gender. And we’ll also look at where people are based, because as David said, people in different parts of the country will buy different products.

So thank you for listening, and I’ll see you next time on The Joy of Marketing

Reader Interactions

Leave a Reply

Your email address will not be published.