Mais conteúdo relacionado
Semelhante a Case Study 2 (20)
Case Study 2
- 1. The Challenge
Does this sound familiar? As a marketing strategist, you are leading a meeting with
a room full of data specialists, data scientists and financial analysts, designing a
project plan to create/update the marketing engine. Up for discussion is an
overwhelming array of data files, tables and links, as well as mountains of external
data. We can’t help but think the marketers in the ‘mad men’ era were so primitive.
Speculating in ideas whereas we are sophisticated practioners that manipulate and
navigate oceans of information to identify and customize, uniquely relevant
communications to each individual customer or prospect. Just look at all the data
we have access to that they didn’t:
ü Complete records of the entire engagement and relationship of every
customer
ü Augmented customer data with a virtually endless stream of external data
including social interactions
ü Complex statistical models are created and maintained by PhD’s, sift through
millions of individuals data elements, predicting future behavior at different
levels of statistical confidence
Despite the allure of executing pinpoint marketing that defines modern era
marketing, we are stuck at the transom discussing how to get started. “How much
data do we really need? Intuitively, we all know that some data/programs will have
an overweighted impact on results, how do we prioritize and determine a cutoff?
How much IT work will be required to modify data to import? How many joins
among tables will be required, and how are the tables refreshed? How many
statistical models are required to narrow clusters, and apply micro-targeting? If the
clusters narrowed become too small, how do we measure performance metrics
through control group hold outs?”
We lie to ourselves to justify that all marketing should be data driven, we can pull in
everything, and let the data itself tell us what we need to know. But we also realize
that senior management is expecting outstanding results, at an unrealistically short
timeline, and ridiculously low costs. We know that we really don’t need all the data.
Plus, as consumers, we receive a regular stream of direct mail and our personal
email boxes are clogged with junk that appears devoid of any relevance. Their
mantra appears to be simply send to everyone, those who need this offer will
respond and every one else will simply ignore it. Amazing. You wonder to yourself
how Lexus or BMW can afford to send out millions of pieces of junk mail or pointless
emails. You wonder who lost the discussion whether to execute a data driven
campaign versus a blanket generic mailing.
The Lies We Tell Ourselves
- 2. Its maddening, as marketing professionals, who are comfortable with the challenge
of crafting a strategy aligned with the data from the marketing engine, and hitting
certain performance metrics on a scorecard tied to payback from a business case.
The challenge is to construct a marketing engine that delivers the financial
objectives while satisfying internal expectations on speed and efficiency.
Why else would the CEO of one of the most influential companies in the world get on
stage, and insist
“We should message a business just the way you would message a friend”
The Process
There is no one perfect way to create and manage a marketing engine. It’s a
combination of science, art, intuition and diplomacy. Constant communication is the
one absolute, communicating continuously across all levels and players.
1. Segmenting
The usual first step is to attain agreement to import/modify/connect as many data
fields as you can afford within budget and timeline constraints. Frequently,
predictive modeling will help parse out critical drivers versus filler data. Trials and
testing using sample extracts can provide early directional insight. Time can be
saved eliminating marginal data elements. A cadence of small scale trials allow you
to start with a larger range of data sets with the expectation that some will fall away.
Another option is to have the data scientists design multiple predictive models each
targeted to achieve a unique outcome. Testing can measure the difference achieved
by triangulating scores from multiple models to improve engagement as well as re-
define communications within the micro-targeting. One advantage of this approach
is that you can determine incremental financial benefit at different tier levels, ie
intersection of 3 models versus only 2 models, or only 1 predictive model.
Applying any combination of model results
will reduce quantities, but should lead to
better personalization of communications.
Scorecard metrics and ROI will determine the
relative effectiveness of personalization at
such a granular level.
And not to be overlooked, relevant messaging
contributes to long term brand value, even if a
customer or prospect does not engage
immediately.
- 5. Reactive tools could be a menu of ‘next best actions’ that align with the segment, as
well as lifecycle management. The stumbling block is the operational interface
delivering a consistent experience across different channels.
4. And Then There Are Random Events
As noted earlier, designing and operating a data driven strategy through a
marketing engine is part art and part science. Certain events don’t comply with all
the wizardry. Here is a typical question, Why did Alan buy a Snickers Bar?
Persona for Alan:
Ø 35 years old and married
Ø Lives in a suburban ranch home
Ø Bachelor degree in Accounting
Ø Likes peanuts, chocolate, pretzels and Doritos
Ø Maintains an active, athletic lifestyle
Ø Drives a Honda Accord
Ø Targeting to retire by 60
Is there anything that segmentation or content strategy would have influenced Alan
to purchase a Snickers bar?
Alan, who rarely snacks, felt hungry, walked over to the vending machine, and
purchased a Snickers bar. When asked how that purchase occurred, Alan replied
Experienced marketing managers recognize there will always be a stream of actions
outside the scope of a marketing engine. Success recognizes what is controllable,
and is not distracted by actions outside their scope. The lure is for less experienced
marketers to become too data dependent, and fail to understand the delicate
balance between controllable and random actions.
The Outcome
There is only one universal measure of success, and that’s the ROI. Creating,
launching and managing a long term strategy through a marketing engine is a
platform for success. Sadly, too many data driven initiatives fail to live up to the
hype and expectations. Marketers oversimplify by assuming you pour data into a
pot, data geeks extract insight, and voila, you have an instant strategy!
that the Snickers bar was located in the
top row at his eye level in the vending
machine, was the first item that caught
his attention, and without considering
other options, dropped his money in the
slot and walked away with a Snickers bar.
- 6. Long term winning, and sustainable performance requires commitment to a
sensible balance among the science, art, intuition and diplomacy. Headwinds may
come from impatient senior decision makers, disruptive events from competitors,
even a lack of commitment from critical internal partners. An experienced data
marketer has effectively dealt with those challenges, and yet delivered critical
results, consistently.
If the strategy is solid, the business case is realistic in delivering the ROI expected,
then patience, confidence and skillful communications can be the difference
between a leader and merely a player.
For more information contact:
Erik LaPrade
https://www.linkedin.com/in/eriklaprade
eriklaprade@gmail.com
cell: (913) 319-9757