Why Tesco Clubcard is a dead end
Once upon a time, a water clock was one
of the best ways telling the time, if you didn’t want to, or couldn’t, estimate
the time by the heavens. Water clocks were exquisite pieces of engineering:
complex, sophisticated and expensive.
You can see a picture of one water clock,
here. It was built by So Sung in China in 1088
and stood 30 feet tall, with five front panels with doors to inspect manikins
which told you the hour of the day and other important times.
Yet despite their awesome ingenuity,
water clocks were an evolutionary dead end. The nature of the process – using
drips of water as the timing mechanism – created certain insurmountable limits
which meant, for example, that the clock would always be extremely expensive to
manufacture, could never get passed a certain degree of accuracy and could
never become truly portable.
If we return to our discussions about
managing relationships with customers, and using customer data, we have a
living example of another splendid dead-end in front of our eyes. It’s called
Tesco Clubcard.
There is no nay-saying it, Clubcard is an
awesome, phenomenon; a superb success which demonstrates oh so much about the potential value and uses of customer data. But
it’s not, and never can be, a prototype or way forward for anything but a tiny
handful of companies. Let’s see why.
First, we have to look at the nature and
the qualities of the data it collects.
– - – It’s collected very frequently, because people shop for food at least once a week
and often two or three times a week. This means it’s possible to identify
trends quickly. It also means that loyalty rewards accumulate in a
satisfying way. So, unlike many other suppliers, Tesco is top of mind, and
people have good reason to update it of basic bits of data like ‘I’ve
moved home’.
- – - It has a very high share of purse. Most shoppers have a main store brand they buy
from, and often 80 percent or more of the food they buy is from that shop.
So even though Clubcard still gets a partial picture of the individual’s
total food purchases, it’s not as fragmented as brands in other
industries. (In some categories such as cars or pensions, you might have
100 percent share of purse but very low frequency of purchase which
drastically limits what you can do with the data.)
- – - It sells a vast range
of items, so not only are people shopping weekly or twice weekly, they are
buying an average of 40 items a time. This means its ability to look for,
and find, correlations and patterns is much, much greater than for most
other brands where there are only one or two items in the shopping basket.
- – - The things people buy are pregnant with meaning,
saying an awful lot about their lifestyles and even values. You know
whether they have pets or children, what their attitude towards the
environment and social responsibility are (do they buy green, organic or
fair trade), how well off they are, and so on.
Each one of these attributes is
important. In combination they are very
important. When you have multiple, frequent purchases that represent a high
‘share or purse’ or ‘share of requirement’ in a category that is central to
people’s lives, the fatal flaws arising from the first law of customer data
(poor data synchronisation), the second law of customer data (inability to
create a genuine single customer view), and the third law of customer data (the
limitations of ‘rear view mirror’ data) still hurt, but they don’t hurt half as
much.
So that is one reason why Tesco Clubcard
is different. If we were to place it on the statistical normal curve in terms
of the nature and type of data it collects it would sit way out on the limb –
the extreme 0.1 percent of all cases: not a good benchmark for comparison. With
the best will in the world, most other marketers (including other types of
retailer) can never emulate the richness or reach of Clubcard. If you try, all
you are likely to do is to incur the expense without the benefits.
But Clubcard’s uniqueness doesn’t stop
there. Tesco is also unique in the degree to which it has operational control.
As a retailer controlling its many interfaces (store formats, online, direct
communications such as Clubcard mailings, traditional advertising) Tesco has a
cornucopia of ‘tweak points’ (price points, shelf adjacencies, different types
of promotion, shelf positions, store locations, etc, etc) that it can
experiment with, in low key, low cost ways – where it has immediate feedback to
its actions.
This combination of a controlled
environment plus immediate feedback (thanks to high frequency of purchase) in
turn, means that it is free of most of the noise and delay that makes
disentangling cause from effect so hard for most marketers. This rapid-fire,
test-and-learn improvement strategy has been a key factor in Tesco’s success –
but it’s an option that’s not been available to most other marketers.
This, in turn, underlines another thing
that’s different about Clubcard. The original theory behind White Elephant or
Full Monte CRM was that the data the organisation collected would provide it
with deep insight into the attributes, preferences and propensities of each
individual customer, thereby allowing it to customise and personalise all
marketing efforts (offers, communications and so on) thereby reducing costs
while maximising returns.
Tesco does a tiny, tiny amount of this: personalising the mailings it sends to
Clubcard members. But in reality this is an extremely thin piece of personalised
icing on a completely impersonal cake. Most uses of Clubcard card data have
nothing to do with understanding individual customers and nothing to do with
marketing communications. They are about aggregate data (patterns, trends)
applied to core operational improvements (‘which promotions work best, under
what circumstances’, ‘what is the most profitable price difference between
private label and the main brand?’). This has got nothing to do with what most
people mean when they talk about ‘CRM’.
This, by the way, is why other companies
with very rich data struggle with CRM. You could argue, for example, that banks
know more about their customers than Tesco does. By looking at your current
account they know virtually everything there is to know about your life. But
(leaving aside the fact that your current account does not provide a complete
picture of your financial circumstances), they cannot do much with this data a)
for privacy reasons and b) because this data doesn’t tell them anything about
how to improve the efficiency or effectiveness of their operations. Knowing
that I spent £80 in a restaurant last week doesn’t help the bank improve its
delivery of banking services. But knowing how I shop for baked beans helps Tesco
improve its delivery of retail services.
This is a simple point but it’s easy to
forget, especially if you think of CRM in terms of ‘better targeting’.
But there’s a fourth way in which
Clubcard is unique. The costs of running an operation like this are
cripplingly, astronomically high. But that doesn’t matter one jot if you have
hundreds of captive suppliers who more or less have to buy your data from you,
thereby turning your cost centre into a profit centre. Tesco generates positive
cash flow from selling Clubcard data to its suppliers – something most other
CRM practitioners can only dream of doing.
So if we put these four dimensions of
uniqueness together …
- 1. an outlier in terms of data richness and accuracy
- 2. unique degrees of operational control
- 3. unique ability to undertake rapid, low cost, low risk
test-and-learn experiments - 4. a profit centre rather than a cost centre
… how many other businesses have the
opportunity to do one of these, never mind all four in combination?
Well, there is an answer, and dunnhumby
have discovered it. Now, I’ve never worked for the business so I don’t know for
sure. But observing from the outside, it looked at though,in its early days,
dunnhumby fell for what still is a common fallacy – the belief that Tesco
Clubcard is an excellent exemplar of a set of universal principles that every
business can apply. So they tried very hard to sell their skills and
capabilities to everybody else: financial services providers, media owners, pharma
companies, white and brown goods manufacturers, etc. In fact, in those early
days, the only companies they didn’t
try to sell their wares to were other grocery retailers, because they
represented a potential competitive threat to Tesco – a possible conflict of
interest.
In fact, as it turns out, other grocery
retailers are the only place where dunnhumby can effectively sell its approach – because that’s the only place
where it really works. Talking in evolutionary terms, Clubcard has evolved to fit
a unique niche and cannot flourish outside of that niche. It’s back to that
waterclock.
Clubcard is indeed a splendid thing. We
have to admire it for its sophistication, ingenuity, beauty and power, just as
we admire Chinese waterclocks. Impressive as they were however, waterclocks
were not the evolutionary route to better ways to tell the time. They were a
magnificent dead end. And so is Clubcard. An evolutionary cul-de-sac.
That’s a great pity because, in a funny
sort of way, Clubcard has done immense damage. Lauded by this that and every
other guru as the poster child of CRM, ‘proof’ that it works, the case study to
emulate, hundreds of would-be CRM pioneers have fallen for the knee-jerk
reaction “Clubcard is a great success, copy it” only to be lured into their
commercial graves.
That’s why we have to be prepared to
think afresh, to be creative, and to look more carefully at what is happening
in the world of customer data around us, including politics and policy. This
world is changing very fast and moving in new and surprising directions, as my blog
After CRM suggests.
If we really want to build better
relationships with customers it’s these new, emerging opportunities that we
should be exploring.
Alan
Mitchell www.ctrl-shift.co.uk
Newsletter
Haymarket © 1957-2013