Guest Comment: Media optimisation in multi-channel retailing

Long gone are the days when direct marketing meant a customer chose something out of a catalogue and phoned in an order. Today, a customer receives several catalogues, emails and telemarketing calls from an apparel retailer. That customer may then search the web for a specific item, and ultimately decide to place an order online. Said differently, the relationship between media and channel has evolved from one-to-one to many-to-many.

Multi-channel marketers are still in the infancy stage of truly understanding the dynamics between media and channels, and in ultimately optimising their investment across all media in order to maximise total revenue across all available channels. While organisational silos continue to be a great obstacle, under-leveraging of analytics is a large contributor to why media optimisation is a very difficult endeavour for retailers.

Listed here are key analytical techniques that multi-channel retailers can leverage in order to measure and optimise spend across media:

Channel migration — helps retailers look at the number of customers and their behaviour across channels (call centre, web, store). Keeping a scorecard of the number of customers and the percentage of their actions in each channel helps multi-channel retailers better align and allocate investment across each media.

Regional market tests — Regional tests are based on a structured in-market testing technique aimed at determining the interaction between two mediums. For example, a retailer can send a specific customer group in a geographical market the full volume of DM/catalogue and email campaigns. Meanwhile, it sends a different group of customers in the same geographical market the same level of DM/catalogue campaigns but throttles back the email campaigns. The test provides retailers with a good read on the reinforcement or cannibalisation effect of email on DM. The results will ultimately help retailers to synchronise and optimise communication plans across media.

Response allocation/attribution — Compares customers’ order activity to their promotion history across all media (email, direct mail, telemarketing, search) to determine if the order can be attributed to a specific campaign, or fractionally to multiple campaigns. Response allocation is a great tool to help retailers measure campaign performance across campaigns based on specific business rules for how to allocate the revenues.

Incremental lift modelling — This is an analytical technique that has been gaining popularity. The intent is to identify customers who are so engaged with a brand they’re very likely to continue purchasing. A company can decide not to promote to these individuals or to use low cost media (email) to promote to them. Particularly useful to retailers that have a pervasive brand presence (large number of stores and large investment in broadcast and general advertising).

Media mix modelling — This is econometric modelling applied to examining how marketing and advertising stimuli affect a retailer’s sales. It uses simulation techniques and is based on past history. Key uses include predicting future sales and isolating the effects of each media. Implementing these models can be costly and time consuming. However, this technique can provide an excellent guide to help companies allocate media budgets at the highest strategic levels.

Media optimisation is an organisational discipline. Unless companies are willing to embrace such discipline, and be willing to change their organisational cultures, structure and processes, media optimisation will continue to be an elusive goal.

• Ian Hitt is the managing director of multi-channel marketing services specialists Epsilon International.