Long-term benefits of using display and sponsored ads

By: Daniela Yu, Sr. Analytics and Media Manager

Advertisers often observe the short-term effects of their ads during activation. But many wonder, what happens afterwards? Our analysis shows long-term sales benefits from display and synergy with sponsored ads. Advertisers can benefit from these long-term and synergy benefits by using display and sponsored ads at the same time.

Story highlights:

Advertisers often are comfortable evaluating the short-term effects of single ad products, but are unsure about the long-term effects and the synergy among ad products. To reveal such effects, analysis has to allow for sales effects after the activation period, as well as consider both the separate, individual effect of each ad product and their additional joint impact (the "1 + 1 = 3" of synergy).

We performed such analysis using data over 77 weeks across the categories of beauty, grocery, health, and personal care. The studied ad products are display impressions and sponsored ads, which includes both Sponsored Brands and Sponsored Products. Nine of the studied brands are leading their category, while 6 are challenger brands (brands that are not category leaders, but are focusing on downstream results to gain share).

1. Display drives sales beyond activation period

When the studied brands invested in display, their sales increased not just in the weeks of activation, but also in the following weeks.

In the visual below, brands gained 100 (indexed) sales during the week of campaign activation. This is how much higher that sales are in the week of campaign activation compared to baseline sales (sales prediction when no campaign). In the next week, after the campaign was over, sales were still 70 higher than the pre-campaign baseline (indexed as 0), and they continued to be higher than baseline for several weeks afterwards.

This long-term effect was substantial: it amounts to 4x the short-term effect for display, and 2x the short-term effect for sponsored ads, which is consistent with advertisers’ perception of display as reaching audiences in the upper-funnel, providing sales benefits that last longer and inform consumer behaviors further down the funnel.

Graph of sales lift over 12 weeks. Brands gained 100 (indexed) sales during week 1 (week of campaign activation); in week 2 sales are at 70; in the following weeks sales continued to be higher than baseline (indexed as 0).

Y axis: Sales lift

X axis: Weeks (Week 1 = week of campaign)

2. Display often shows synergy with sponsored ads

Synergy is significant for about half of the analyzed brands. Their sales benefits from using display and sponsored ads together go beyond the sum of the separate, individual effects of each ad product. The remaining brands still obtained the separate sales gains from each ad product, but didn’t obtain synergy effects.

The visual below shows how this works for a specific brand example. In weeks when both display and sponsored ads support were in the bottom half of all weeks, sales were on average 100 (indexed baseline). Average sales were 12% higher in weeks when display support was high in the top half, and 34% higher in weeks when sponsored ads support was high in the top half. However, in weeks when both are high in the top half, average sales were not 46% (12% + 34%), but 78% higher than baseline. Therefore, this brand got an extra of 32% (78% - 46%) while using display and sponsored ads together, beyond what it did by using them separately.

Synergy effect of sponsored ads and display on retail

100%
112%
134%
178%

3. Challenger brands are most likely to enjoy synergy

Digging deeper into which types of brands show synergy, we uncovered that it is most common among challenger brands. In fact, 4 out of 5 studied challenger brands show significant synergy. The likely explanation for this is that more shoppers are unaware of challenger brands (upper-funnel) initially, which means that lower-funnel actions are more likely to help motivate them to purchase.

In shoppers’ path to purchase, advertisers can use display to drive awareness and interest in a brand or product, and then help drive purchases with sponsored ads. Finally, the studied challenger brands also showed higher long-term sales effects for Sponsored Display.

The visual below shows the average sales effect for the studied leading brands (in blue), and the studied challenger brands (in orange), respectively. While the typical leading brand experienced most of its sales benefit in the week of the campaign (as in visual 1, above), the typical challenger brand experienced most of its sales benefit 1-2 weeks after the campaign.

Percentage of advertisers who adopted Sponsored Brands

Percentage graph of the average sales effect for Challenger brands vs Leading brands over 7 weeks. Leading brands experienced most of its sales benefit in week 1 (the week of the campaign); challenger brand experienced most of its sales benefit in week 2-3 (1-2 weeks after the campaign).

Weeks (Week 1 = Week of increase in display spending)

Challenger brands

Leading brands

As Warren Buffet, the "Oracle of Omaha" advises, the key is to "Always invest in the long-term." Here is a clear method of following that advice using Amazon ad products.

Methodology

This analysis uses data from 14 brands. Data for each brand comprise 77 weekly observations (a total of 1078 observations) from January 2018 to July 2019. The 14 brands cover 6 large categories from 3 verticals (e.g., Nutrition and Wellness and Baby Care in Health and Personal Care, Coffee and Cold Beverages in Grocery, and Mass Skin Care and Mass Cosmetics in Beauty). The 14 brands include 9 leaders and 5 challengers. We define challengers to be brands that are not category leaders, but are focusing on downstream results to gain share. For each brand, the analysis aims to explain sales as a function of display impressions, sponsored ad impressions, their interaction (to measure synergy), retail impressions, monthly seasonal control variables, and tentpole control variables. The number of weeks each impression variable drives sales is revealed by vector autoregressive models, which quantify the over-time response of sales (visuals 1 and 3) as a function of changes to advertising impressions from display, sponsored ads and their interaction (synergy).