December 30, 2013 at 14:51Add to Cart vs. Buy in One Click: AB Test Results
Over the course of a month, we conducted an AB test on the site of one of our customers (an online gift store), in which we compared the conversion of visitors into paid orders for three different options for placing an order:
Immediately on the main thing: option 2 won with a conversion increase of + 58% in comparison with option No. 1. Under the cut, more about the test, option No. 3, some numbers and graphs.
When a new user visited the site, one of three order options was randomly opened with the same probability (all options are shown on KDPV):
No. 1. Only the "Add to Cart" button.
No. 2. “Add to Cart” (main order option) + “Buy in 1 click” (additional option).
No. 3. “Buy in 1 click” (main order option) + “Add to cart” (additional option).
In the session for the buyer, the option number assigned to him was kept, so that when going to other pages of the store, the buyer saw the same order option, and not every time a new one. A custom variable with the option number (_setCustomVar) was added to the Google Analytics tracking code, and based on the value of this variable, segments of site visitors were formed for each of options No. 1, No. 2, No. 3.
When buying using the "Add to Cart" button, the buyer can collect a lot of goods in the basket and then place an order in the classical way in three steps: choosing a delivery method, choosing a payment method, entering a delivery address.
When buying in one click, only the phone and the name of the buyer are filled in (something like “call me back”). The address and amount of delivery the operator of the online store will clarify later by phone.
The first test - all three options No. 1, No. 2, No. 3 - was conducted from November 17 to December 10, 2013. In the chart below, you should pay attention to the Total Events indicator of the options - this is the number of placed orders in each of the three options (counted by order placed events):
No. 1 (basket only): 14 orders
No. 2 (basket + 1 click): 26 orders
No. 3 (1 click + basket): 22 orders
Comparing options No. 1 and No. 2 and taking into account only paid orders (these numbers, for obvious reasons, are slightly different from those on the graph), the addition of the “Buy in 1 click” button allowed to increase conversion by 58% (before the start of the test, the indicator of the total conversion of all site visitors to payment The volume of orders was 0.59% and in option No. 2 it increased to 0.93%) The average order amount in option No. 2, however, decreased by 7% compared with No. 1.
Obvious was the victory of options No. 2 and No. 3 (with the possibility of buying in one click), so option No. 1 was removed, and only No. 2 and No. 3 continued to participate in the competition. In the second part of the test, from December 11 to 28, 2013, option No. 3 showed a greater result:
No. 2 (basket + 1 click): 43 orders
No. 3 (1 click + basket): 69 orders
According to preliminary results, option No. 3 (1 click + in basket) won, but after dropping out all abandoned orders and calculating the amount of sales, option No. 2 still became the most effective for this project. Firstly, due to the fact that 1 click orders had about three times more failures(every fifth order in one click was failed: customers simply make test orders, enter “1234” instead of a real phone number, etc.; for a classic order, the number of failed orders is about 7%). Secondly, the average cost of a paid order in option No. 3 was significantly - by 35% - less than in option No. 2.
A rather noticeable increase in orders in 1 click for option No. 3 in the second part of the test compared to the first part was observed partly due to seasonality (New Year), partly due to the fact that the online store published paid reviews (ads) in some entertainment sites and blogs advertising specific product positions - there were more orders in 1 click on conversions from such posts than through the basket (here the numbers vary depending on the product, therefore are not given; on average, for all advertising transitions, the 1-click option was one and a half times more used).
There is practically no pronounced difference in preferences in choosing an order in 1 click or through the basket either by region, or by browser, or by OS, or by user device (just in case, we checked all the options), and the most obvious dependence is observed by the traffic source. The most common case is when referring from search engines to pages of specific products. If a person knows what he is looking for and gets directly from the search engine to the page for a specific product, then most likely he uses the 1-click order option (here, too, the data varies significantly depending on the product item).
In terms of total sales, option No. 2 showed a result 13% higher than option No. 3.
The task of the test was to calculate how much conversion could be increased at the working online store of a relatively small size by simply introducing the possibility of express order. Since the test was conducted not so long, it would be fair to round the result of increasing the conversion for this project to + 50%. The same estimate is generally true for general sales.
For online stores in which a typical order consists of one product, it is probably better to use the one-click order option as the main option. In this example, the average quantity of goods in one order is 1.8 (and if we consider additional services to goods, for example, gift wrapping, then 2.6), and option No. 2 worked better with a basket as the main way to place an order.
Perhaps in a number of projects it is worth using a combined scheme in which to use the 1-click purchase option as a priority for selected products, which are usually ordered in one copy (this analysis was not performed in this test).
The online store " Mega Gifts " shared the experience , all data was published with his permission. The test was conducted by Webasyst staff. The store operates on the basis of the Shop-Script 5 engine .
You can read about the technical setup of such an AB test for the Shop-Script 5 engine here .
Happy New Year! We wish you a productive year 2014!
Immediately on the main thing: option 2 won with a conversion increase of + 58% in comparison with option No. 1. Under the cut, more about the test, option No. 3, some numbers and graphs.
How was the test conducted?
When a new user visited the site, one of three order options was randomly opened with the same probability (all options are shown on KDPV):
No. 1. Only the "Add to Cart" button.
No. 2. “Add to Cart” (main order option) + “Buy in 1 click” (additional option).
No. 3. “Buy in 1 click” (main order option) + “Add to cart” (additional option).
In the session for the buyer, the option number assigned to him was kept, so that when going to other pages of the store, the buyer saw the same order option, and not every time a new one. A custom variable with the option number (_setCustomVar) was added to the Google Analytics tracking code, and based on the value of this variable, segments of site visitors were formed for each of options No. 1, No. 2, No. 3.
When buying using the "Add to Cart" button, the buyer can collect a lot of goods in the basket and then place an order in the classical way in three steps: choosing a delivery method, choosing a payment method, entering a delivery address.
When buying in one click, only the phone and the name of the buyer are filled in (something like “call me back”). The address and amount of delivery the operator of the online store will clarify later by phone.
results
First test
The first test - all three options No. 1, No. 2, No. 3 - was conducted from November 17 to December 10, 2013. In the chart below, you should pay attention to the Total Events indicator of the options - this is the number of placed orders in each of the three options (counted by order placed events):
No. 1 (basket only): 14 orders
No. 2 (basket + 1 click): 26 orders
No. 3 (1 click + basket): 22 orders
Comparing options No. 1 and No. 2 and taking into account only paid orders (these numbers, for obvious reasons, are slightly different from those on the graph), the addition of the “Buy in 1 click” button allowed to increase conversion by 58% (before the start of the test, the indicator of the total conversion of all site visitors to payment The volume of orders was 0.59% and in option No. 2 it increased to 0.93%) The average order amount in option No. 2, however, decreased by 7% compared with No. 1.
Second test
Obvious was the victory of options No. 2 and No. 3 (with the possibility of buying in one click), so option No. 1 was removed, and only No. 2 and No. 3 continued to participate in the competition. In the second part of the test, from December 11 to 28, 2013, option No. 3 showed a greater result:
No. 2 (basket + 1 click): 43 orders
No. 3 (1 click + basket): 69 orders
According to preliminary results, option No. 3 (1 click + in basket) won, but after dropping out all abandoned orders and calculating the amount of sales, option No. 2 still became the most effective for this project. Firstly, due to the fact that 1 click orders had about three times more failures(every fifth order in one click was failed: customers simply make test orders, enter “1234” instead of a real phone number, etc.; for a classic order, the number of failed orders is about 7%). Secondly, the average cost of a paid order in option No. 3 was significantly - by 35% - less than in option No. 2.
About option number 3
A rather noticeable increase in orders in 1 click for option No. 3 in the second part of the test compared to the first part was observed partly due to seasonality (New Year), partly due to the fact that the online store published paid reviews (ads) in some entertainment sites and blogs advertising specific product positions - there were more orders in 1 click on conversions from such posts than through the basket (here the numbers vary depending on the product, therefore are not given; on average, for all advertising transitions, the 1-click option was one and a half times more used).
There is practically no pronounced difference in preferences in choosing an order in 1 click or through the basket either by region, or by browser, or by OS, or by user device (just in case, we checked all the options), and the most obvious dependence is observed by the traffic source. The most common case is when referring from search engines to pages of specific products. If a person knows what he is looking for and gets directly from the search engine to the page for a specific product, then most likely he uses the 1-click order option (here, too, the data varies significantly depending on the product item).
In terms of total sales, option No. 2 showed a result 13% higher than option No. 3.
conclusions
The task of the test was to calculate how much conversion could be increased at the working online store of a relatively small size by simply introducing the possibility of express order. Since the test was conducted not so long, it would be fair to round the result of increasing the conversion for this project to + 50%. The same estimate is generally true for general sales.
For online stores in which a typical order consists of one product, it is probably better to use the one-click order option as the main option. In this example, the average quantity of goods in one order is 1.8 (and if we consider additional services to goods, for example, gift wrapping, then 2.6), and option No. 2 worked better with a basket as the main way to place an order.
Perhaps in a number of projects it is worth using a combined scheme in which to use the 1-click purchase option as a priority for selected products, which are usually ordered in one copy (this analysis was not performed in this test).
Thanks
The online store " Mega Gifts " shared the experience , all data was published with his permission. The test was conducted by Webasyst staff. The store operates on the basis of the Shop-Script 5 engine .
You can read about the technical setup of such an AB test for the Shop-Script 5 engine here .
Happy New Year! We wish you a productive year 2014!
Only registered users can participate in the survey. Please come in.
What ordering option do you usually use in online stores?
- 49% Regular Basket 203
- 6% One click purchase 25
- 12.3% Both one and the other 51
- 3.8% I use a basket, but prefer a one-click order 16
- 28.7% The main thing is that the checkout process is simple and short 119
