5 initial steps to measure the impact of online advertising on offline revenues

Advertisers in industries with expensive products which are often bought offline, such as luxury products and cars, often face the following question by their management:

“Now that we are advertising digital and we can track customers, show me how these efforts pay off for the business.”

Now technically it is possible to measure and track the customers, but it’s often not done. Measuring the payoff, ROI, ROAS or CPA in an actionable way can unleash exponential growth in any organization, given of course the product being sold having the right product-market fit. 

In this article, we’re showing 5 initial steps to measure online advertising to offline revenue impact. As a prerequisite, we strongly recommend having at least 6 months of data both on the spendings and the revenue side.

1. Get access to all your costs data
   

   Extract all the data from all the places where you buy media and extract them on a breakdown by channel by day. If you advertise in multiple markets, it’s good to have the country as an additional dimension. Usually, the CPA by country varies greatly.
   
   Here the source platforms will be Google, Facebook, Bing, Twitter, Linkedin, etc..

    

   

   Cross channel costs

2. Get access to all your revenue data

   Now we want to get access to all the revenue data across all channels where you have your revenues. Here it is again important to have the channel (e.g. offline store, e-commerce stores, retail partners, online store). 
   Here your source systems will be your CRM system where you track all revenues.
   As for the spendings, the more granular the data, the better. It’s best to have daily data by offline revenue source. If you operate in multiple countries we strongly recommend to also have the data on a country level.

   Note: This step often would require help from the finance team. It’s important to clearly explain to them what you do and get their buy-in.

    

   

   Offline Revenues

3. Match your data on a daily, weekly and monthly level

   Now depending on our technology stack, we take all the data and we put them into a database table or a simple Google Sheet or Microsoft Excel file

4. Analyse your data

   Now depending on our technology stack, we take all the data and we put them into a database table or a simple Google Sheet or Microsoft Excel file. The best way to look at the data is by creating a simple bubble chart with the following setup:

   Series: All conversions you have through the chosen period
   Entity: Week (or Day or Month, this will need experimentation on your end and depends on the product you are selling)
   X-axis: Online Spending
   Y-axis: Offline revenue
   Bubble size: ROI

   The results will be a bubble chart similar to what you see below:

   

   Now you can start identifying trends and you should be able to see patterns which help you to investigate further.

5. Refine your analysis

   Now update the bubble chart and specifically to (a) limit data to a period which is expected to have a common seasonal pattern (to reduce seasonality effects), and (b) adjust revenues using a price index, if it is available (which should take into account product mix and promotion events). We’ll cover more on the topic of building your media mix model soon.

Wrapping it up

To drive results you will need to

1. 1. Get access to all the data and merge it into one single place
   2. Understand your control variables (seasonality, price events, …)
   3. Take action on the insights and iterate

Last but not least, our platform helps clients to get insights as described above and makes sure the data stays fresh (no more CSV and Excel data wrangling). More information and our pricing can be found here.