The search budget comes with a clear online sales case, while in-store sales sit in a separate reporting system. Before approving it, decide whether the plan credits in-store sales affected by search advertising. The board case still needs a margin test because offline-only return on ad spend averaged below break-even, while several retailers' estimates cleared it.
Broad category search advertising can create in-store sales that online attribution misses.
Data chart
Percent rise in total-store sales over matched control stores, across each test plus the two weeks after. Total store is the whole store in some experiments and a group of departments that includes the advertised categories in others. One bar per retailer experiment; the dashed line is the combined estimate for all 14 experiments, not an experiment.
Every one of the 14 retailer experiments that reported total-store sales shows an estimated rise, so losses in the other categories each experiment measured did not cancel the gains in the advertised categories.
Model estimates from 14 randomized market-level field experiments at well-known U.S. multichannel retailers; a 15th experiment did not report total-store sales. The model pulls each experiment toward the overall average; the retailers' own reported rises ran from 0.01% to 8.80%. Pale bars marked (n.s.) have ranges that include zero. Broad category terms were held in search positions 1-3.
Data chart
Dollars of extra in-store sales for each extra dollar spent on search ads, one bar per retailer experiment; the dashed lines mark break-even at a 35% gross margin (2.86) and the combined estimate for all 14 experiments (2.50).
The combined in-store return on search advertising falls short of break-even at a 35% gross margin, while seven experiments' estimated returns sit above it, each with a range that reaches below break-even.
Model estimates of return on ad spend from the same 14 experiments, counting incremental brick-and-mortar sales only; online and future sales were excluded, so the returns are conservative. The model pulls each experiment toward the overall average. Pale bars marked (n.s.) have ranges that include zero; that test compares each return with zero, not with break-even. Each of the seven estimates above break-even has a range that reaches below it.
Key takeaway
Broad category search advertising can create net new in-store sales that online attribution misses.
The experiments covered large U.S. multichannel retailers advertising on broad, non-branded Google search terms in positions 1 to 3. They measured store sales and offline return on ad spend, not online sales or sales beyond two weeks after each test.
Source
Kalyanam, K., McAteer, J., Marek, J., Hodges, J., & Lin, L. (2018). Cross-channel effects of search engine advertising on brick & mortar retail sales. Quantitative Marketing and Economics. Experiments ran on Google.com. The protocol committed in advance that neither Google nor the retailers could drop an experiment from the analysis on the strength of its result, and non-significant experiments were retained. https://doi.org/10.1007/s11129-017-9188-7