Page 11 - AeM_Apr_2017
P. 11
RESEARCH, ANALYSIS & TRENDS
An all-encompassing attribution model
is still utopian
In general, companies use marketing attribution software With their spreadsheets and their simple last-click mod-
to determine how their marketing activities and cam- els, the danger is that marketers ignore this warning
paigns contribute to sales revenue during the prospect- and wait too long for a solution that will never be deliv-
ing and sales processes. Getting the desired insights, ered. Instead, the message from the experts is ‘don’t let
however, is often easier said than done, since a closed the perfect be the enemy of the good.’ Attribution is not
sale is normally the result of a combination of several about being 100% right all the time. There are still great
factors and conditions. Hence, it’s not easy to identify leaps to be made simply by being ‘less wrong’.”
what really tips the scale for making a purchase.
So, if the machines won’t do all the heavy lifting, deliv-
For that reason, marketing attribution software is usually ering the best outcomes will ultimately depend on the
a hub software, pulling in data from other marketing and skills of the marketers and business analysts, along
software tools. It commonly integrates with software so- with the help they receive from agencies and vendors—
lutions used by sales, marketing and public relations to and of course their platforms.
helps marketers identify which channels or campaign
tactics deliver the best return on investment and to tell There is, for instance, a well-arranged list of the top
what works and what doesn’t. attribution advertising companies on Thalamus’ web-
site that could be useful, another one is provided by
With today’s digital technologies becoming more sophis- VentureRadar or DataFox, just to provide a few, where
ticated by the day, and the emergence of machine learn- the best Attribution Software vendors are usually deter-
ing and artificial intelligence, some believe that a ‘grand mined by customer satisfaction (based on user re-
unified attribution model’, that encompasses all online views) and scale (based on market share, vendor size,
and offline data, will soon become a reality. For others, and social impact).
like AdRoll’s Managing Director APAC, Ben Sharp, how-
ever, “given attribution is more an art than a science”, Considering that attribution is key for marketers looking
making staff’s experience critical. to optimize media channels in today’s increasingly frag-
mented world, it came as a surprise to read that adop-
In the company’s whitepaper Attribution: Building a Busi- tion of attribution has been slow. But that’s what the
ness Case for Attribution, he points out as well not to fall research provider eMarketer revealed in one of its lat-
for the common misconception that machines can make est reports, even lists a few reasons why that is so,
the decisions, stating, “programmatic advertising has such as lack of knowledge or experience or time and
created a sense that decision-making can be outsourced money issues.
entirely to the software. Supply- and demand-side plat-
forms, retargeting and, lately, data management plat- Anyway, marketers at least seem to have finally started
forms have all added to the growing expectation around to put a spotlight on attribution, as surveys from the
algorithmic perfection. But in the real world of attribution Interactive Advertising Bureau (IAB) Data Center of
as it sits today, it might be time to reassess this view. Excellence and the Winterberry Group, or the program-
The message from practitioners is that an all- matic advertising platform Rocket Fuel, reveal. We’ll
encompassing attribution model based on one grand see how the trend will gain momentum in the future. ◊
unifying theory of everything will never actually exist. By Daniela La Marca
MediaBUZZ Pte Ltd - Independant ePublisher for Asia