Fresh Data Blog
Fresh Data Archive
The Evolution of Attribution
Date: June, 2013 --
Technology is beginning to generate tools to enable the
marketing world to more effectively and accurately answer the age-old question.
It’s the question every merchant since the first trade of a stone arrowhead for
a stone knife over a wood fire has wanted to know. What made this person engage
with me? What caused them to trade? Or, more contemporaneously, what medium, promotion,
and /or ad placement drove, either wholly or fractionally, the purchasing
decision of any given customer?
great copywriter Denny Hatch ‘attributes’ action by a prospect reading an offer
to one of the basic human instincts: Fear, Greed, Guilt, Anger, Exclusivity,
Salvation or Flattery. What influence
has caused a person to do something? If
it’s one week to Mother’s Day, good copy in a letter addressed to men with an
offer of flowers and candy might be premised on fear, attributing fear of
domestic punishment as a motivator of action.
And while this is
undoubtedly true for selling copy, the current subject of “attribution” is many
times more complex and starts with the problem of channel choice, not the
advertiser’s choice, but the customer’s.
In a nutshell, the goal of
any attribution strategy is the discovery of how to turn prospects into
customers as economically as possible in an “omni-channel-media” world. In a
direct mail-only world, this was relatively simple. What do your customers look
like and what do they respond to at what time of year? In a media-over-loaded
marketing environment, it is many times more complex. The goal is to generate
insights into customer behavior and reaction to impressions and, when said
customer makes a purchase, attributing credit to one or more of the sources of
these impressions. Seasoned marketers have learned to fully appreciate that the
“last-click” metric is quite possibly very misleading.
What are the impact of paid
ads, search, email, mobile, social media and database-founded efforts? What
does each contribute on the path of the buyer from unknown to prospect to
customer? How much effort/money should we allocate to each medium? At what point in the person’s trip to us
should we use each medium? And how much should we spend on each step? Knowing
the total picture will prevent cutting programs that were the true driver of
sales, despite not being the last impression logged or medium/outlet used in
the purchase. We’ve all heard stories of retailers who saw huge spikes in their
eCommerce sales, decided to cut back on print catalogs to move those dollars
online, which then led to a significant drop in online purchases because, as
the retailer came to realize, it was the direct mail catalog that was driving
website visits and purchases.
At the beginning of our
experience of the Internet as a commercial space, as demonstrated in the
aforementioned example, the retail and catalog companies who ventured online
quickly discovered the complexity of attributing media to purchase behavior. Many
readers will recall this type of fervent discussion about what amount of
catalog buying was driven by the website, how much by the catalog and vice
versa. And recall the fights between the catalog department and the new
“website” / eCommerce department on who had “really” sold the goods! You
couldn’t return a product you’d bought on the website at the local branch of
that store, and you couldn’t pick up the product you had purchased online at
the local branch, either!
These seem like elementary
questions from the dark ages. There were few if any tools to do the
Now the data exists, or can
be captured, manipulated, and analyzed with quantative analysis tools as well
as direct and fractional attribution models. These will help in understanding
the path to purchase and thus in allocating budgets to relevant media.
These tools can disclose the
time from first contact to conversion and help shorten it. Perhaps presenting a
different ad on the second appearance of the prospect will do that. If we know
their path to their first click to the website, their similarity to existing
customer populations helps inform what we show them on arrival. The optimal
amount of time we present a banner or sidebar ad to a particular prospect can
be predicted from analyzing behavior of similar prospects.
A frequent question is the
significance of the channel through which the prospect first discovers you. The
media entry point may be critical to a particular population, less so for another.
The more data we have, the more accurate the prediction of behavior can be, the
higher the ROI.
Implementing these programs
is potentially disruptive. Having selected a vendor, you will have to assess
your goals, assets, infrastructure and resources, both human and machine. Decisions must be made as to what data must be
collected, where it will come from and how it will be imported. Data must then
be aggregated and normalized. Personnel may need retraining, and marketing and
product silos may have to be collapsed. There will follow a testing phase which
begins the process of providing customer behavior insights which lead to
improved ROI. With each data-input cycle these insights, and the range of
topics addressable, should broaden.
In short, the machine will
“learn” and make more and more accurate predictions on marketing strategy and
Whether you’re a season
veteran with tried and true processes or still finding your way, talk to the
direct marketing experts at Data Services, Inc. about setting up the right
processes and platforms that will allow you to better manage your direct
marketing attribution strategy.