Anticipating consumers’ every move to conversion is essential to realizing a strong return on marketing investment (MROI), yet, pervasive methods for understanding consumer behaviors fall short in a multichannel environment.
In the digital world, channel savvy, highly mobile, multiple device-happy consumers are in control. While they move from one media channel to another—and from one device to another—companies across industries are struggling to keep pace.
Anticipating consumers’ every move to conversion—and measuring every interaction across every channel—is essential to realizing a strong MROI and to sustaining competitive differentiation. Yet, pervasive methods for understanding consumer behaviors fall short in a multichannel environment.
As such, various attribution models and approaches have emerged to help businesses better capture and analyze consumers’ purchase journeys across both offline and online channels. Though a singular approach for measuring consumer data has yet to emerge, multichannel attribution holds exciting promise.
December 6, 2012
Altimeter: The 2014 state of digital transformation June 2014
Accenture: Multichannel Attribution Dec 2012
1. Accenture Interactive | Point of View Series
Multichannel
Attribution
Measuring Marketing ROI in the
Digital Era
2. Multichannel Attribution
Measuring Marketing ROI in the Digital Era
Digital technologies have A typical consumer journey is anything but • How did a particular sale happen?
linear, single channel or reliant upon one
empowered today’s consumers to • Who should get the credit for it?
device. It may start in front of the TV and
get what they want, when they progress toward a tablet, then switch to a • ow much credit should be attributed to
H
want and where they want. So smart phone, make pit stops at a product each consumer interaction across channels,
while consumers may have few website followed by a social network site, and on what basis?
problems moving from one media for sharing news about the product, and end
with a purchase at a store. See consumer • ow should the investment be apportioned
H
channel to another and from journey scenario. across channels?
one device to another, business
is having a hard time keeping up In order to serve this channel-savvy, highly
with them.
mobile, multidevice-happy consumer, chief These questions
Anticipating the next move in a
marketing officers (CMOs) have their work
cut out for them. These CMOs need to have have become the
consumer’s journey to conversion,
an accurate understanding of consumers—
their intentions, impressions of products subject of some
and measuring every interaction
across channels, is a significant
and services and their behavior; to pinpoint
exactly which marketing channels—online or
lively discussions
challenge. Pervasive methods offline—are yielding maximum MROI. on direct, last
in use today for associating
consumer behavior with marketing
For example, knowing the impact of paid
search engine marketing (SEM), online
action or last click
investment fall short of capturing display media, natural search marketing
based on search engine optimization (SEO) or
versus multichannel
the contributions from multiple
channels on a specific path to
affiliate partners, on both online and offline attribution.
customer behavior. Or understanding how
purchase. As marketing leaders email, mobile-optimized websites, online
shift budgets from channel to video ads, social media, mobile display ads
and the like, work together with offline
channel, new approaches are media. For that, CMOs must have accurate
required that provide better insight answers to questions such as:
and ultimately drive smarter budget
allocation and improved marketing
return on investment (MROI).
3. A typical consumer journey scenario
Dave is sitting with a tablet watching a show on Dave does a Google Search on his tablet, which pops up a
1 TV. During one of the ad breaks he sees a discount
2 paid search ad. He clicks on the ad and goes to the phone
commercial for mobile phones. manufacturers website, which lists not only additional
discounts and insurance plans, but also has links to reviews.
Dave sends the link to his email address for access later.
On his way to work Dave opens the email on his smart Later, Sarah logs into Facebook on her new phone and
3 phone and forwards the link to his friend, Sarah, who is
4 shares her news about the purchase with her friends.
looking for a good deal on a mobile phone. Sarah goes Some friends search for the phone retailer and make a web
to the phone manufacturers website on her laptop, reads purchase. Some put the item in the cart then abandon the
reviews, then drives to the nearest store to purchase the digital journey – preferring to drive to their nearest store to
smart phone. try the phone and complete the purchase.
4. Last action dilemma Lately, however, it is being widely
One of the biggest problems today is that, acknowledged within the industry that
while consumers operate in a multichannel attributing the cause of a conversion
world, the metrics of their interactions primarily to the last action often leads
are still siloed. That is, we have metrics to overattribution of consumer response
for measuring consumer activity in to that action, ignoring other influences,
each channel but not for measuring the such as a TV advertisement or online paid
entire consumer journey in an integrated search campaign. This is resulting in the
manner. The need to capture and analyze emergence of econometric MROI analyses
this journey as it takes place across both which focus on, measure, and analyze
offline and online channels has led to the historical data of all known interactions
emergence of various attribution models over a specific period of time to in order to
and approaches. derive multichannel attribution.
For the most part, direct attribution has Regardless of which approach one
been a popular approach for analyzing embraces, it is absolutely clear that unless
consumer data. Not only because consumer data is captured in real-time,
alternative approaches are just beginning across channels and devices, any exercise
to emerge, but also because direct that gives credit to consumer action per
attribution is simple to implement. It calls specific channel or device may lead to an
for attributing conversion (or purchase) to inaccurate interpretation of results and
the “last action,” such as a mouse click or thereby to undesirable business decisions.
direct mail response.
5. Multichannel attribution Dynamic tools and capability • eural Networks: Leverage machine
N
Analyzing consumer behavior is getting There is no “one-size-fits-all” approach learning algorithms that use complex,
more complex, not only because of the sheer to successful multichannel attribution. nonlinear mapping functions for
volume of data, but also because that data The consumer data landscape and business estimation and classification. Neural
comes from multiple sources, including environment are highly complex and ever- networks prove their worth in
consumer touchpoints, data providers, and changing. Therefore, high performance, multichannel attribution where there
advertisers. So how to accurately measure dynamic tools and capability are vital is often a lack of consistent historical
and interpret consumer data? to success. The techniques required information or an absence of a
may include: theoretical framework around causal
Though one reliable approach for measuring relationships among variables.
consumer data that is good for all situations • Clustering: Group customers with similar
has yet to emerge, there are a number of traits and profiles, and use clustering Integrated analytics
multichannel attribution approaches that to identify homogeneous patterns. While multichannel attribution can yield
are yielding positive results. There are Each attribute within each cluster is hitherto unobtainable levels of granularity,
three fundamental aspects to successful allocated a specific weight to identify it is not a panacea. It should be aligned
multichannel attribution. its relative importance to the cluster, to, and integrated with, other analytics
and across clusters. This is not a once- techniques as and when needed. For
Big data mindset and-done exercise, but a technique that example, if analytic processes are already
Data needs to be gathered and analyzed helps dynamically identify trends and in place focusing on both media and
from across consumer touchpoints. To similarities in customer behavior. marketing mix, as well as investment
achieve that goal, analytics engines that optimization, then adopting an integrated
can process a huge volume of data and drill • Logistic Regression: Isolate and measure
analytics approach is a powerful
down to various levels of data granularity the impact of the multitude of media, combination. This type of approach will
are required—a must for understanding marketing and service influences on allow companies to apply marketing mix
both individual behavior and the behavior customer behavior. Through scenario modelling to understand the marketing
of a segment of consumers across channels. planning, optimize future activities to effectiveness at a regional level, and
This requires robust data architectures drive sales, reduce churn and improve conduct multichannel attribution analysis
for enabling systematic consumer data customer satisfaction. to dive deeper into the data for insights
acquisition, processing, population, and surrounding media, customer experience
reporting, as well as data analytics. and customer behavior.
6. Comparison of Direct and Multichannel Attribution – an example
To demonstrate the difference between direct and multichannel Taking a direct attribution approach, the Google paid search
attribution, consider for a moment the following example. A ad will be identified as the activity that influenced the sale.
consumer sees a TV ad that results in him wanting to purchase The multichannel attribution approach however, considers
a specific product. He goes online to make the purchase and the influence of both the TV and the Google paid search ads.
decides to use his Google search engine to find the product URL. See table below.
His Google search returns a paid link promoting the product
which the consumer clicks on to take him to the site where he
makes a purchase.
Direct attribution Multichannel attribution
Summary of Uses the information provided at point of Recognizes that multiple inputs drive a purchase
approach purchase, whereby a click or entry of a code and quantifies, or attributes, value to each input.
indicates where to attribute the purchase or
connection.
Scenario results The purchase or connection is attributed to The purchase or connection would likely be
google keyword search. attributed to both the tv and google paid search
advertisements, and in a way which allows the
diminishing return on increasing investment in each
activity to be captured—much more reflective of
true customer behavior.
The direct attribution approach has the activities drive base consumer behavior decision to cut TV advertising and direct
advantage of being simple to calculate, in a statistical sample, while others drive more funds to paid search, which could be
providing top line metrics for marketers incremental consumer behavior. the result of the direct attribution approach,
to justify investment and manage costs. the value of TV advertising is recognized
However, direct attribution also leads The multichannel attribution approach when considering future investment and
to linear thinking—the assumption that provides greater insight into the data funds are not redirected in such a “black and
everything has a fixed cost per connection or influencing the purchase or connection. In white” response. Multichannel attribution
click, and if marketing investment is doubled, this example, attributing the influence to enables optimization of forward looking
the consumer response rate will also double. the TV ad on the purchase will impact MROI budgets, identification of cost savings and
In practice, there is significant variation calculations of both the TV and Google improved MROI.
in effective cost per response, as certain paid search activities. Instead of taking the
7. Driving Consumer Relevance consumers or to a specific segment of 1 A phenomenon, we call “the R Factor” or “consumer
relevance at scale.” For a full discussion on the R
The channel-savvy, highly mobile, consumers at scale. Companies will be Factor, see Baiju Shah and Nandini Nayak, “Got
multidevice-happy consumer has able to “test the water” for new products the R Factor: Driving breakthrough performance in
the Era of Relevance,” http://www.accenture.com/
created a new playing field for marketing and services, learn from early consumer SiteCollectionDocuments/PDF/Accenture-Relevance-
organizations. Today, the effectiveness of responses, and make the necessary changes At-Scale-POV-WEB-5April.pdf accessed June 28, 2012.
marketing strategy is increasingly measured before launching them to a wider audience.
in terms of its relevance to the intent and As more attention is focused on the
changing preferences of consumers across last-action dilemma, we are likely to see
channels and touchpoints—at both macro multichannel attribution becoming a key
and micro levels of granularity.1 component of not just marketing strategy,
but of a company’s overall business strategy.
As the practice of multichannel attribution
gains popularity and attains a certain level
of maturity, companies will be able to deliver To learn more about developing
the right message, via the right channel, a multichannel attribution
at the right time to a larger audience, as
approach, contact:
well as make accurate investment decisions
across channels and touchpoints. The Conor McGovern
way companies run marketing campaigns conor.mcgovern@accenture.com
should change for the better—with data
granularity and new analytics techniques,
marketing organizations can access micro-
segmentation data and deliver targeted
marketing campaigns to individual