People use the Internet in funny ways. Ever looked over somebody's shoulder when they perform a search? Ninety percent of the time (I made this number up, but it's a lot) you'll think that they're searching in a stupid way.
Some people will put in an incredibly short search term, and just trust the search engine to know what they want. This is the kind of person who, if asked to look for a pair of red stilettos, would type "shoes" into Google.
Other people will use very long and specific search phrases, hoping that it'll crop up somewhere in exactly that format. "Where can I buy Beyonce's red shoes that she wore at the oscars?" was a phrase that sticks out in my memory from an AdWords search query report (we got the sale though!).
Most people are somewhere in between, but chances are that if you ask somebody to look for something, you won't agree with the way they go about it.
It's similar when trying to look at browsing patterns over Christmas. Here I'm going to present some basic stats about what happened to UK AdWords traffic over Christmas this year, but trying to extract helpful information from it is anything but basic.
We've taken AdWords data from 250 websites across a cross-section of industries, all based primarily in the UK. Websites range from large (millions of visits per day) to small (thousands of visits per week) but we've excluded the very largest so that the data doesn't represent just a few websites. We also removed any websites with incredibly large changes in traffic volumes (e.g., those that grow a lot between these periods) since they would influence on period more than another.
This data doesn't break down websites by industry sector or demographic (keep an eye out for a future article) but instead is intended to see what happens to overall user activity over the Christmas period.
This graph shows the relative total AdWords click volumes over the 2010 and 2011 Christmas periods. 2011 volumes are higher every day in this period, but come very close in early December and the final few days before Christmas. Those of you in the UK may remember a very cold period in 2010 that affected overall Internet traffic quite considerably: many people couldn't get in to work, and many ecommerce websites had to stop fulfilling pre-Christmas orders due to delivery problems.
The most obvious things to note are that in 2010 the increase in traffic over Christmas didn't happen until Boxing Day, but in 2011 Christmas Day traffic picked up a lot and was significantly higher than Christmas Eve.
Of more interest is how the traffic at Christmas behaved relative to the rest of the year. We took the average daily clicks for the final three months of each year, and set that value to 1 in each case. We then plotted each year's data (for the same period as before) against that three month average index. We got a very different story.
2010 overall had a more significant Christmas than 2011, but that wasn't true all the time. These numbers are more difficult to interpret because of the effect of weekends falling on different dates, but the final week of November and first week of December were clearly much more important in 2010 (relative to the rest of the final three months of the year) than they were in 2011. Our Christmas Day increase in 2011 now looks less like a big rise, and more like an unusually low traffic day on Christmas Eve.
What can we do with this? On its own we can't do much.
This high level data doesn't help us to work out much about our websites or what we should do to plan for Christmas next year. This is only going to give us information about how the populace at large is behaving and interacting with AdWords ads by day.
We can have some idea how many people are searching on each day, but that doesn't let us know which days will be quiet for our own campaigns. As I'm sure you can guess, an e-commerce campaign for consumer electronics would perform pretty differently to a B2B campaign.
Your next steps are to try to work out what your own stats were for your website, and what you can learn.
- Download your AdWords stats for each day, including impressions and impression share.
- Divide impressions by impression share to get an idea of search volume that day.
- Assuming you didn't make major adjustments to pricing in that time period, add your CTRs and conversion rates to this mix.
You should now have an idea how many potential conversions are out there over the Christmas period, based on your own website data. Compare against the baselines on this chart and you should know whether you should be planning for a big Christmas or a small one.