58% of marketers aren’t satisfied with existing attribution models for digital ads

Anne Freier

In Mobile Advertising. August 12, 2019

Digital marketers are not happy with existing attribution solutions, according to a survey by QueryClick, the digital marketing agency.

The survey among 700 respondents from across retail, travel and finance industries found that 58% of marketers feel restricted by their current attribution models which do not allow for long-term payback implementation.

Sixty percent of respondents said they felt pressured to over-invest in paid search (PPC) because of its instant results and measurability. Yet, marketers admit that these strategies aren’t aligned to maximise return on investment.

Meanwhile, agencies are emphasising PPC precisely because it allows them to show off the immediate benefits to their clients.

The sources of pressure among respondents include budget (26%), management pressure to focus on a certain channel (22%), agency decisions (16%) and marketer decisions (16%).

“It’s particularly discomfiting because one of the most attractive aspects of digital marketing is its measurability. For management to disregard that measurability is illogical, and probably a bit dispiriting for those in their teams who will often feel that they are being ignored,” said Chris Liversidge CEO and founder of QueryClick.

Over half (59%) of respondents also agree that Last Click attribution is a gross oversimplification as an attribution option.

Marketers tend to use a broad range of attribution models including data (21.5%), position-based (20.5%), first interaction (12.5%) and last interaction (11.5%).

However, despite the abundance in attribution options, 58% of marketers feel restricted by current models.

“One plausible explanation is that attribution is regarded simply as a historical reporting tool, recording what investments are believed to have yielded ‘yesterday’, and that attribution insights that should help to invest more efficiently ‘tomorrow’ are considered to be of little accuracy or value. But this explanation raises a further question: if the historical record is apparently of little use in predicting the future, then to what extent is that historical record an accurate one? And how on earth is it being validated if not with predictive ability?” Liversidge added.

For 61.5% of respondents investment changes made based on attribution insights fail to deliver the predicted results. At the same time, 78.5% have confidence in their current attribution systems accurately measuring the impact of TV and radio on their marketing channels.

When it comes to programmatic, 80% of respondents are already investing in programmatic ads; 56% plan to increase their investment over the next 12 months.

However, ad fraud continues to be a core issue among marketers. Liversidge suggests:

“As estimated by respondents, the mean proportion of programmatic ads seen by humans is a measly 44.06% – considerably less than half. The single most popular response was ‘20%-50% [of ads are seen by humans]’, suggesting the industry has a long way to go before ad fraud is remotely considered to be eradicated.”