Brand Equity Attribution Modelling (BEAM) is a program created as a result of a partnership between PHD and PropertyGuru, with the goal of achieving a quantum leap in media effectiveness by challenging industry best practices.
Shedding the norm of measuring brand equity on an interval basis, PropertyGuru adopted a continuous measurement approach while building media plans on the basis of high-velocity experiments.
The marketing team used quantitative techniques to stress test and discover the most effective media mix and adopted a performance-based agency compensation model for PHD.
“By applying the principles, we have been able to increase media effectiveness by 20% to 30% per market,” said Bjorn Sprengers, the chief marketing officer of PropertyGuru. “Simply put, it means we can grow our brand equity with 20% to 30% per market at stable budgets, even more, if we add investments.”
According to Sprengers, the BEAM regression model works by correlating media channels to a Brand Query Index (BQI) which offers a very high ‘goodness of fit’ whereby the coefficient of determination in Malaysia is over 85%.
“Per channel, we now understand how much we need to invest to get a maximum yielding impact on brand equity,” said Sprengers. “We express these investment levels in response curves that show beyond what level of investment, impact on BQI start to diminish.”
This is followed by the calculation of the optimal media plan on basis of the best combination of channels plus investment levels. Comparing these optimums with the historic cost per BQI-point then allows PropertyGuru to prove the delivery of efficiency by 20% to 30% while cutting 50% of brand spend wastage.
According to Sprengers, the BEAM approach allowed PropertyGuru to become in market leader in 2017, with 0.9 times the SEO share-of-voice of the incumbent in the first quarter and 1.3 times by the fourth quarter.
“We have observed that 25% of this result came from BEAM optimisation,” Sprengers said. “The logic is that a stronger brand leads to stronger performance in branded keyword sets. Particularly in online property in which a big chunk of organic traffic is driven by branded searches, investment in brand helps to drive organic traffic."
Given that brands typically struggle with top-of-funnel attribution, the objective of BEAM was to find an evidence-based model towards enhancing up-funnel Marketing ROI.
The project was fueled by Sprengers desire to hold marketers accountable for business impact and the reality that bottom-funnel ROI attribution is easier to achieve as most media investments are digital.
Of the lessons learned by Sprengers's team using BEAM, the outliers were that the BEAM regression model delivered a high coefficient of determination while identifying channels that performed and those that did not.
Of the former, content and brand partnership channels were the best performing, while BEAM proved that each media channel has its own unique profile of optimal spending levels whereby marketers may experience diminishing marginal returns.
Sprengers emphasised that marketers need to understand the response curves of media channels in order to optimize media effectiveness towards BQI.
Running the regression model under BEAM twice a year, Sprengers anticipates that the biggest optimisation potential for future iterations will lie in better understanding the impact of investment in a specific combination of channels as opposed to investment in individual channels.
For advertisers and agencies interested in adopting the BEAM approach and model, Sprengers recommends a drastic culture shift towards bringing accountability into the equation.
“I’d say the genesis of this change is also embedded in the way we believe and do things at PropertyGuru,” said Sprengers. “We work with a growth mindset, where we are unafraid to question conventions and experiment newer solutions that can make us more impactful.”
Sprengers believes that the BEAM model has widespread applications for marketers across industries and verticals, with an easily adaptable methodology for digital consumer brands and marketplaces, adding that the specific model coefficients and associated marketing plans will differ.