Multi-Channel Attribution Modelling Helps a Retail Industry Giant Achieve 5% Growth in Sales
Client: eCommerce business unit of a leading retail industry client specializing in grocery, furniture, building and home trade, sports goods and apparels in the United States. The global retail industry includes establishments selling commodities for household consumption. Today, amid varying consumers’ preferences and the rising consumer spending habits, the growing concern for healthier lifestyles is […]READ MORE >>
Client: eCommerce business unit of a leading retail industry client specializing in grocery, furniture, building and home trade, sports goods and apparels in the United States.
The global retail industry includes establishments selling commodities for household consumption. Today, amid varying consumers’ preferences and the rising consumer spending habits, the growing concern for healthier lifestyles is encouraging the growth of the retail industry. Moreover, retail industry customers are craving for authenticity, creativity, and convenience in the products being offered owing to the recent shift to a more customer-driven economy. However, a few factors are expected to influence the growth projections of the retail industry in the coming years.
- Increase in online marketing and retailing initiatives: Owing to the relentless growth of technology, customers in the retail industry are looking for convenience while purchasing products. Additionally, a shift toward seamless and agiler online marketing initiatives are expected to address the purchasing discrepancies among the customers.
- Understanding the customer: Understanding the behaviour of the customer has become an issue for the retail industry as buying habits change dynamically, subjected to various uncontrollable factors. As a result, it is crucial for retail firms to keep up with customers and to serve their ever-evolving needs to be competitive.
These factors are forcing businesses across the retail industry to leverage the use of multi-channel attribution modelling solutions. The multi-channel attribution modelling solutions help businesses determine the marketing channels and campaigns contributing the most towards driving leads and ultimately sales. These solutions also help companies improve the way they are allocating their marketing dollars and adequately fund all the channels driving purchases.
The Business Challenge
A leading retail industry client in the US wanted to optimize the ad budget for digital channels to ensure that they were investing in revenue generating channels and align campaigns to the right media at the right time. Their biggest challenge was to understand the real path to purchase. This was due to the fragmented online environment with more than three agencies for different channels including social, online video, paid search, and display ads. As a result, it was difficult to determine which media channels and tactics were driving sales. Additionally, the retail industry client needed a more specific attribution based approach for digital channels and address business predicaments.
- The client provided data from multiple agencies for seven different channels
- Finalized business scope and consumption use cases through client discussions
- Cleansed the data based on business rules
- Integrated the current and historical data at the individual customer level
- Aggregated the data at the DMA level
- Tagged the data for each URL / source based on channel tagging rules for attribution model
- Analyzed the clickstream trends, calculated click-through rates (by DMA, State, Website, etc.)
- Univariate analysis to analyze the impact of each variable on final activity (sales orders)
- Multi-variate regression and Hidden Markov models to attribute sales across multiple channels
- Designed dashboards on Tableau to ensure clickstream trends, web metrics and sales could be tracked
Insights and Business Benefits for the Client
- The bulk of the budget was divided between social and search channels (paid and organic) based on the click-through rates
- The client was spending disproportionately on social channels since large % of sales were attributed to them due to which they had a better reach on social media
- Quantzig proved that the paid search channels were responsible for a greater share of the revenues (per order) hence the budget needed to be re-allocated
- Based on the results of the attribution model, Quantzig recommended
- Spend for each channel
- Expected response from each channel
- Expected revenues
- Based on the client’s current requirements and advertising strategy we suggested that channel attribution be performed at DMA level
- Helped provide clear comparison of test and control regions within the same states
- Identified the most significant variables affecting sales conversions – geography, DMA, channel (organic search, paid search, social channels), etc., creative attributes (size of display ads, pre-roll vs post-roll ads on YouTube, etc.)
- The dashboard allowed the client to understand the recommended budget by each channel for each DMA
- Helped the decision makers provide guidance to each media team for achieving targets at a more granular level, i.e., # no. of impressions from the display channel