Proof Is In the Pudding: Impact of Online Advertising on In-store Sales

ASAP, EOD, POV, FYI, ETA, COB, SOV, ROI – when did a few letters begin to mean so much? Out of all the everyday acronyms thrown around, ROI is by far the most important metric for our clients in today’s market. The key question from every one of our clients is “What is my return on investment?” For years we, as a search company, have been able to point to high online sales ROI for our clients, but an element in the ROI equation has been MIA – how can we prove that online advertising (both search and display) actually affects in-store sales? If you take into account consumers researching prices and specific products online and going in-store to purchase, the overall ROI for search marketing drastically increases. But how can we prove this online to offline phenomenon is occurring?

For years small research tests and anecdotal stories pointed to online assisting offline, but more in terms of brand awareness and consideration. Our team wanted to prove online advertising contributed to offline sales. Not top of the funnel activities like others have proved, but actual hard number sales. 

One of our clients appeared to be the perfect fit for this type of test. With more than 1,000 retail locations nationwide and no e-commerce shopping available (except for gift card purchases), we could create a test to measure the impact of online advertising on in-store sales. In order for this test to work, we would need the client to share sales, unit and transaction data, as well as have our research team identify suitable test and control DMAs for search and content advertising. We pitched the idea to the client and their online media agencies as a test to compare the sales lift produced by online advertising in comparison to other media vehicles in the marketing mix as a more cost-efficient means of assisting all stores, including struggling locations. The existing media plan for the client relied heavily on television commercials, newspaper inserts, shared mail and direct mail, with a small percentage of the existing budget going to online and the largest percentage going to direct mail. If we could prove a strong ROI for online, while showcasing additional benefits of online compared to traditional advertising (very quick turnaround times, minimal production costs, and being present when consumers were actively engaging for relevant content) we hoped we could make a case for a shift in online's importance in the media mix.

The client and agency both agreed that the test appeared compelling and interesting. With their interest piqued, our research team moved forward, creating test DMAs to heavy up search and content advertising in specific product categories, especially furniture. Furniture is typically a high consideration purchase, with a long purchase cycle and everyone wanted to better understand how online advertising impacts the buying cycle.  Control markets were also established where the selected product categories would not be promoted and all other advertising remained consistent across all DMAs to ensure the marketing mix was not interrupted. The next important step was that the client agreed to share its sales, unit, and transaction data for specific time periods (one month before the testing period, the three week test period, and two weeks after the test, for both 2007 and 2008) for the purposes of test design and  analysis. Ultimately, 59 DMAs were selected as test markets and 93 DMAs were selected as control markets for the test, focusing on specific furniture products and the client’s brand terms.

Throughout the test (conducted in late September/early October), our team had its concerns – the economy was crumbling before our eyes, retailers were declaring bankruptcy, and consumer confidence was near an all-time low – was this the right time to go about this huge test with our client? We anxiously awaited the results with antacids in hand, and were pleasantly surprised by the final results.

Despite the poor state of the economy, low consumer confidence and decreased sales across the board for national retailers, our client saw a 2% overall sales lift in the test DMAs over the control markets, which is on par with results from pricey newspaper inserts and other offline media. Specifically, the furniture category saw a sales lift of 1.4%, but that number could have been higher if we had continued to analyze sales data for a longer period of time (the sales cycle for furniture extends longer than the five weeks of data we were able to collect from the client). Additionally, great lifts were seen in categories such as dining and decorative accessories (2.9% and 2.2% sales lift respectively) that have shorter sales cycles and could have resulted from furniture searches driving foot traffic in-store. Struggling stores saw an even more impressive lift of up to 5.3% in revenue – well above what we thought possible in certain markets.  And back to the all-important three letters of R O I – our test delivered a 300% ROI for our client, with $3 in sales for every $1 spent online. 

Excited, ecstatic, thrilled, ready to get to work on building out more keywords and campaigns – that is where our team is today, and proud that we could partner to help our client to be more efficient with their marketing spend to drive more sales for their business.

Posted by Kelly Wall, Google Retail Team