Improving Ecommerce Retention and Revenue with Personalization
Imagine you go back to a shop you have been before, and the salesperson recognizes you. He also remembers that you were looking for a green jacket, size M.
Not only that, he knows what you like, so he suggests that you combine the green jacket with a particular green suit that goes with it.
Chances are, you’ll keep going back to that store, because you now have a personal connection with the salesclerk. He knows exactly what you want.
Personalization in ecommerce is like that. It’s using technology to personalize, in scale, the experience of every customer of an online store, like a good salesperson would in a retail shop.
It’s not surprising to learn that most consumers like online stores that have some type of personalization engine. Nearly 60% of online shoppers think it’s easier to find more interesting products on a personalized online retail store and 56% are more likely to return to a site that recommends products.
This preference for personalization has a significant impact on revenue. A survey conducted by Econsultancy in association with Monetate, shows that ecommerce companies using personalization technologies are seeing, on average, an uplift of 19% in sales.
Choice Overload and Personalization
Personalization works because it’s an antidote to a phenomenon called “choice overload.” According to a study conducted by Iyengar and Lepper, too many product options can lead to people feeling overwhelmed and then dissatisfied with whatever choice they make. This is especially true for online stores, since there isn’t a limited number of products that can “fit” in a website, allowing stores to have an unlimited variety of items to sell.
Having a vast catalogue of choice is an advantage because it improves the chances of serving each customer with the products that are right for them. But offering too many choices can sometimes do more harm than good, and prevent customers from finding the right product. The solution is to offer fewer and more relevant options out of the products you have, tailoring the shopping experience to each customer’s needs.
Examples of Personalization
It’s not surprising that large online companies, such as Amazon, Netflix and Spotify have been recommending products on their sites for years.
By cutting down the choices to a reasonable number, Amazon is leading users to make better decisions about what to buy, helping them escape the feeling of choice overload.
Personalization technology is getting more affordable, too. What was once only available to a team of engineers in giant software companies is now within anyone’s reach. This easy access to cheap technology is changing the way small and medium-sized online merchants interact with their clients.
How personalization works
The engineers at LimeSpot, a content attribution and personalization technology company, recommend three basic steps a company needs to take to implement a personalization strategy on their website.
1. Categorization
In the same way that our brains need to organize information in “boxes,” such as movie genres, people’s nationalities, and types of food, personalization systems need to categorize products according to what is common among them. So before implementing any type of recommendations engine, you need to have a well–organized inventory or products in your system.
Better organized products will give recommendations engines (and shoppers) the ability to understand how products relate to each other.
2. Cut down the options
According to LimeSpot, seven seems to be the magic number of recommended products that humans are most comfortable with. Amazon shows more than that (though they do have a uniquely large list of products), but we suggest you use seven as a maximum number to start with.
The principle behind this small number of items is to counter the feeling of choice overload mentioned earlier in the article. When your personalization engine shows fewer and more relevant products, clients have a better experience. Resist the urge to add too many products to your recommended section.
3. Gain trust and convince users to sign up
Once you have shown users that you’re a reputable store, you should try to persuade them to sign up to some kind of membership. You can use incentives such as discounts or a smoother experience in the checkout process to convince them to do so.
When users sign up you will be able to store more information about them, which you can then use to recommend more relevant products.
By using information from their current visit, their previous purchases, and what other people similar to them have bought before, you should be ready to recommend the following options:
- Most popular
This is a simple product recommendation, based on what is popular in your store in the customer’s region.
- Similar items
If your products are well categorized, you should be able to recommend similar products on your product pages. Amazon showed me different models of cameras and accessories below the information about the camera I was interested in buying.
- Recently viewed
If you have an extensive line of products, people may spend a lot of time browsing through them. You can help them go back to what they were looking at before by displaying products they have recently viewed.You can use sign–ups or cookies to remember those choices in future visits, so when customers come back to the site they will see something like this:
- You may like
This recommendation is based on a more sophisticated understanding of the user, using their purchasing history to recommend new items to buy, independently of how they relate.
If your store is not too big, you should be able to recommend most popular and similar items manually, according to your understanding of your own products, and simply add product images on product pages. However, a larger store and a more sophisticated personalization strategy will require a more complex technical implementation of a recommendation algorithm.
The return on investment of a personalization solution is well worth the effort of implementing it, though. An analysis of our partner LimeSpot’s recommendation app (starting from $9.95/month) has shown the following improvements a few months after instalation.
- Increase in Revenues: 29.7% – 41.72%
- Increase in Transaction Value: 5.25% – 6.4%
- Increase in Conversions: 5.56%
If you build it yourself, a great engineer could produce a simple recommendation engine in about two months (at a cost of approximately $30,000), while sophisticated systems such as Amazon’s can cost more than $10M to develop
Consumer behavior doesn’t change very much, whether people are shopping online or offline. Investing in a personalization system is like hiring an intuitive salesperson to help themchoose something from your online store.
