Understanding marketing attribution
Marketing attribution is a means to understand the impact that marketing initiatives make on a purchase, sale or various other KPIs like driving footfall, increasing brand lift and so on. A creatively well planned digital advertising campaign combined with a comprehensive attribution model allows marketers to get a firm handle on marketing analytics that reduce wasted ad impressions. This makes way for sharper, more accurate and pragmatic marketing.
As we explore the value of MTA in the CPG world, it’s helpful to gain a deeper understanding of the most commonly used attribution models by marketing researchers.
- Marketing Mix Modeling (MMM)—Primarily Offline
The MMM is a top-down attribution model that analyzes aggregated historical data, most typically from offline media sources like TV, radio and print. Organizational-level metrics that are used for strategic planning, spending budgets and analyzing performance are usually delivered with a marketing mix model.
The MTA model takes a bottom-up, granular approach requiring user-level data. The focus is on the consumer’s digital journey, looking at the path to conversion across multiple touchpoints and over a time period. MTA measures household-level or individual lift of marketing levers or a combination. It helps marketers evaluate how much credit to assign to each element of a multichannel marketing campaign or each touchpoint. MTA allows measurement of digital touchpoints beyond aggregate channel metrics, such as ad copy or keyword level.
- Unified Measurement Approaches (UMA) – Offline and Online
UMA answers questions that span both the tactical and strategic impacts of marketing. These approaches attempt to resolve the challenges of disparate, unlinked methodologies and insights. It can look not only at the digital touchpoints but also at exposure from owned media programs (CRM) as well as in-store and TV. It can identify and parse out both offline and online efforts.
Data and analytics in marketing attribution
We are looking at several millions of analyzable impressions per ad campaign depending on the type of product being advertised and targeting parameters. Each impression can be also associated with information about the action taken by a user all the way till the purchase of the advertised product, or where she dropped off in the path to purchase. All this translates to large volumes of data that can turn into valuable insights for the marketer. However, on many occasions, the data trail left behind by the customers is left untouched. A report states that upto 73% of all data retained by an organization is unused for analytics. This just goes on to show how crucial it is for brands and ad platforms to realize and leverage the immense potential of data analytics for accurately understanding the performance of marketing campaigns.
Challenges in building marketing measurement models for CPG businesses
CPG companies have the least amount of CRM data as compared to Retail, BFSI, Telco and Travel verticals. In addition to this, a research by CGT reveals that 80% retailers don’t share online customer behaviour data, which is vital for CPG brands to create better marketing campaigns.
Watch Rineet Ratnakar – CIO – Hygiene and Home, North America, Reckitt Benckiser describe some of the challenges they face in attributing the outcomes of marketing spends
Here are some key concerns that a competent Multi Touch Attribution model needs to address to help in improving marketing effectiveness:
- Missing or lack of access to all the relevant data sets that affect a campaign including POS data, and campaign metrics such as impressions, clicks, creative, site, audience level data
- Varying levels of coverage of different data sets. E.g POS at store level, FB at DMA, TV at
- Failure of conventional models in effective attribution
- Top down approaches (regression, clustering) aren’t useful as there are, many causes for variability, delays in gathering sufficient data, absence of time to design and run large experiments and so on.
- Bottom up approaches like (First touch, Last touch and Linear models) need data at lowest granular levels and consumer data – which is only sparingly available to CPG brands
Apart from these, the model also accounts for various other factors influencing Incremental Sales such as seasonality and weather to reduce attribution errors, the impact of multiple Campaigns running at the same time with same audiences, halo and synergy effects.
Scalability: An effective model is also built to process historical data every time the pipeline runs. The model accounts for different types of datasets that would be needed for different types of sales and campaign strategies.
A powerful MTA model falls short in its ability to solve the marketer’s attribution challenges in the absence of intuitive dashboards. Dashboards complete the model’s ability to guide the marketer in the direction where the marketing dollar is performing at its best. A good dashboard helps the marketer with:
- Faster and timely decision making
- Better understanding of marketing tactics & campaign performance
- In depth analysis – slice and dice
- Improved tracking across KPIs
Find out how Sigmoid’s MTA accelerator helps marketers obtain real time analytics for timely decision making.