eCommerce sales figures do and don’t suck on a Friday. This is a real account of a client problem that we scratched our heads around for a while, but solved after getting deep into performance data.
This post explains what the data illustrated, what we analysed and how we worked out why eCommerce sales suck on a Friday.
Friday (and Saturday) eCommerce Sales Suck
eCommerce conversion rates fluctuate from day to day, week to week and season to season as sales do in almost every business. There are innumerable influences on patterns and disruption of patterns of buying, or traffic volumes and of intent and conversion rates, which ultimately influence the numbers we really create about; quantity of sales, cost per conversion and the average order value.
The number of sales and the cost per conversion both pivot around the most important factor of all, Conversion Rate (the percentage of visitors who buy something in your store).
Factors That Influence Your eCommerce Conversion Rate
In no particular order, and not an exhaustive list:
And by quality, we mean high intent traffic.
Search traffic typically converts best as it represents a user, right now searching for what you sell. Within that though, keywords (search terms) represent different stages of the search funnel.
A search for “gifts for girls” is early in the funnel. It’s pretty broad and has low intent. We don’t know what sort of gift the searcher wants yet and the chances are that they don’t either. We also don’t know what their budget is. We call this a discovery keyword. Both of those factors mean a low conversion rate for this search.
A higher intent term like, “chocolate hamper” represents a user we’re much more likely to convert, assuming your store sells chocolate hampers, of course. “cheap chocolate hamper” is even better qualified and we’d expect that to produce a higher conversion rate, again assuming that your store can fulfil the need.
You’d think that it goes without saying, but eCommerce stores miss this point all too often.
If you have more choice on your site, I’m more likely to find a gift, the shoes, the holiday or the hamper that suits my needs. A wider price range means you’ll satisfy more visitors and convert at a higher rate. More variety to choose from, especially with fashion or other purchases where personal taste is a factor, is critical to conversion rate.
Again, pretty obvious. Having the product(s) in stock and available for same day despatch will have a huge impact on your average conversion rate.
Users who have bought from your eCommerce store before and who had a good experience have a higher conversion rate than new customers.
Users who know your brand from Display Ad campaigns or social media engagement, your YouTube channel or even a print ad are much more likely to buy from your eCommerce store than one they’ve never heard of.
Of course, the design, product presentation, navigation and overall User Experience is part of that trust too and also has a significant impact on conversion rates.
We frequently achieve 25-100% increases in conversion rates through improved User Experience.
It isn’t always a factor, but often can be, particularly for traffic arriving through Google Shopping or comparison sites. This is especially important for commodity items like DVDs or ink cartridges – price and customer service are really the only differentiators in that space.
We know from thousands of campaigns and transactions that higher value and more complex products have a longer search funnel and that it is necessary for your ads to be present 100% of the time when users search, or the journey to the sale breaks, lowering the conversion rate and pushing up the cost per conversion (see Impression Share vs Conversion Rate).
Yes, surprise, surprise; you’ll sell more Christmas decorations at Christmas, and swimwear sells better in the spring and summer. But that’s not the end of the story, not by a long shot.
Daily environmental issues can create fluctuations in search behaviour and intent. Some days, you’ll get the same volume of traffic from the same keywords from the same ads to the same products on the same pages at the same price. And the conversion rate will be lower.
If you analyse your conversion data by day of the week over a few months, you’ll see that particular days of the week are better or worse than others. Which days those are will depend on your audience and products, but you will find consistent patterns.
The temperature rising enough for a visit to the beach will reduce traffic, but also reduce the intent users have and therefore your conversion rate.
The World Cup being on the TV had a HUGE impact on a client campaign we worked on. Sales literally flatlined for a completely unrelated product.
School holidays can have a major impact on eCommerce sales too.
What Happens on Fridays?
We’ve already established that days of the week can produce different conversion rates. eCommerce sales figures fluctuate up and down as does the search volume.
This was different though. This was dropping from a 6.6% average conversion rate to 4.4%. Every Friday.
Server Speed. Starting with the quick checks to exclude them from our enquiry, a server being slow every Friday would be a little weird, but isn’t beyond the realms of possibility.
Page load times have a significant impact on conversion rates, particularly on mobile devices.
Page load times fluctuate. There are some peaks and troughs. Nothing correlates with Friday performance.
Ad positions and CTR. A quick check of the AdWords account reveals no correlating data there. Ad positions are consistent, CTR is consistent, traffic volume is down, but not significantly. Even then, it might explain a reduction in the number of transactions, not the conversion rate.
Reference other eCommerce Store Data. Friday’s search volume is lower on Fridays across a number of different eCommerce stores in a number of different vertical markets. The conversion rate is also a little lower, but nothing like the 35% we’re experiencing every Friday, and it’s low on Saturday too, but not quite so bad.
Analyse the Search Funnel. We know that many eCommerce transactions need a number of ad impressions and a number of clicks on average, before they convert. We know that sometimes that means that discovery and shopping start on one day, but can complete several days later. We call those early interactions, Assisted Impressions or Assisted Clicks.
Maybe Friday behaviour is more assist clicks which convert into eCommerce sales on Saturday, Sunday or Monday?
No. After some deep data analysis, we established that this theory was a dead end too.
We Conceded. We set a bid adjustment in Google AdWords to reduce our bids by 50% on Fridays to reduce the losses. It would be better to have no costs and no sales than lose money on every one.
Average Sales Value. But wait. There’s something odd here. Although the conversion rate is 35% lower and the cost per conversion is 50% higher, Friday’s average sale is the second highest of any of the days of the week. Thursday is the highest, if you’re interested.
This makes no sense. Friday’s conversion rate is the lowest of the week by some margin. Friday’s cost per conversion is the highest, again by a margin. But Friday’s average sale value is the second highest for the week.
Delivery. This is a home improvement accessory. It sits in the DIY space. It strikes me that if I’m buying a gizmo on a Friday, I’m going to want to install it over the weekend.
The delivery options on this site are; Free delivery (5 days on average), and paid for options for next day and 2-3 days. Next day delivery is £4.95. I can have my order on Saturday, but it’s going to cost me.
I suggest that the reason conversion rates are poor is that customers are impatient for their goods and then grumble to themselves about paying for delivery. A significant chunk, say 35%, abandon the cart, deciding that they’ll buy online with free next day delivery, or they’ll just drive to the local DIY store and save the delivery charge, also enjoying the peace of mind that they know they’ll have their order on Saturday.
Correlation. It’s a great theory, but we need to correlate some other data to make sure we’re not misreading the data.
We analyse orders with optional delivery fees by day of the week. What we found was that 70% of customers opted for free delivery, but that percentage changed dramatically across the course of the week.
Early in the week, users were less likely to pay for delivery. Mid-week produced the average orders with delivery paid extra. Friday orders are 40% more likely to choose any of the delivery options and Saturday 30% more likely.
Friday eCommerce sales figures suck because customers want their stuff. Friday sales suck because customers are unwilling to pay for next day delivery.
How did we fix it? Please comment below….