Seasonality is one of the most important factors that affect PPC conversion rates in online retail. In fact, the impact of seasonal factors alone is often higher than all the other factors I’ve written about thus far combined.
‘Tis the Season
“Seasonality” is another way to state what most retailers already know: the demand for most products fluctuates throughout the year. Some obvious examples can be seen in the demand for toys spiking as the December holidays approach, demand for costumes spiking as Halloween approaches, and—perhaps less intuitively—demand for wedding supplies spiking in January and June. There are also “micro” seasons that might not be as obvious. It can take some elaborate data slicing-and-dicing to spot these. You have to look across dimensions other than category and sub-category to see these micro trends.
Brick and Mortar retailers understand the importance of promoting the right products at the right time by rearranging their merchandise. It’s just as important to proactively promote the right products online as seasonal peaks approach.
The Impact of Seasonality
Both conversion rates (CVRs) and average order values (AOV) change during seasonal spikes.
Conversion rates often jump dramatically when traffic spikes occur. Adlucent’s research shows increases in CVRs for some categories as high as 300 percent. We believe these drastic changes are caused by customer urgency. For most of the year, demand for products is relatively constant, driven by discretionary purchase decisions. But at certain times of the year, shopping becomes more “hurried” relative to the normal latency that might exist. For the marketer, PPC advertising is more effective when customers are more motivated.
During seasonal spikes, AOVs change, too—often downward. Our experience tells us this is because customers become more focused as they systematically check off gifts for everyone on their list, whereas during the rest of the year they might be looking to buy a book for themselves and be more susceptible to up-sells, cross-sells, and other point-of-sale recommendations. The average price range of gifts—which is potentially less than people would normally spend on themselves—is a factor here, too.
The Importance of Prediction
The dramatic shifts in CVR and AOV during seasonal spikes have a significant impact on how aggressively you should bid in your PPC campaigns. Recall the model I’ve discussed before, which illustrates the relationship between CPC and CVR:
CPC = (AOV × CVR) ÷ ROAS
If your AOV is held constant and your CVR triples, then your CPC can (theoretically) triple while still keeping your ROAS constant. But if you’re adjusting the CPC in response to shifting CVRs and AOVs, you’re too late (see Paid Search Look-Back Periods).
Instead, retailers need to predict seasonal changes and adjust proactively. This involves modeling year-over-year changes in CVR and AOV, normalized by recent performance, then adjusting CPCs and building-out campaigns—all well in advance of the forecasted seasonal spike.
As an example, by modeling year-over-year performance, Adlucent knows to anticipate a spike in traffic and CVR in the DVD category during the Christmas holiday. Specifically, we notice that recently-released DVD series do very well, as do certain types of documentaries. By predicting this spike, Adlucent focuses on comprehensively building out these products proactively in late summer. DVD sales are relatively flat during the first half of the year, so if we relied only on recent performance we would miss this spike. Missing spikes like these is a risk of over-relying on short look-back periods.
“Year-over-year” doesn’t always refer to the same day of the year. Black Friday and Cyber Monday are not on the same calendar date every year, for instance. Getting your dates lined up correctly is important, since customer buying behaviors vary by day of week during the entire year, and conversion rates trend upward during holiday seasons as shipping deadlines approach. Overlaying these factors on top of your predictive model improves your campaign performance.
Harder Than It Sounds
For a retailer with thousands or even millions of products, it’s not a stretch to say that managing seasonality without automation is impossible, because of the challenges we’ve raised here:
Identifying seasons and micro-seasons for every product
Predicting changes in AOV and CVR based on historical sales data, day of week, day of year, and other factors
Adjusting CPCs proactively in anticipation of forecasted changes
Managing seasonality effectively also requires deep expertise in both retail and consumer behavior. Adlucent combines this expertise with our Deep Search software platform to enable proactive bid management that doesn’t just acknowledge seasonal swings but takes advantage of them to maximize sales and ROAS.