In the previous instalments of this four-part series, I’ve laid out a simple guideline for an efficient product recommendations setup on an online store, and I‘ve described how various recommendations widgets work and I’ve explained how purchase attribution works in a product recommendation system.
In the final part I’ll lay out a simple framework to help online e-commerce store owners decide if they need to add personalisation to their store.
The purpose of this write-up here is to offer both a reasoning method and few tangible data points to facilitate the “add personalisation” decision.
A lot has been said about personalisation over the course of the years, and currently, a simple google search for “personalised online shopping experience” yields hundreds of relevant results.
In addition, success stories about how this big brand and that, had used personalisation to help drive their next billion in sales do come up in industry news headlines. The term “Personalisation” has thus, become ubiquitous on the web.
Since you are reading this, I’m assuming you know what it means and that you are interested in maybe employing personalisation on your online e-commerce store.
First things first, if you are just starting out, don’t focus on personalisation. The time for it will come later, and paying attention to other things will help your store stay in the black at the beginning of its journey.
A good time to start thinking about adding personalisation is when your online store reaches 50K unique organic visits a month.
One other thing to be clear about. The framework shown here is not strictly related to product recommendations and can be used when deciding if the online store needs other types of personalisation, be it personalised content, personalised category pages, personalised checkout experience, etc.
This guide, however, should not be applied to personalising email campaigns, as factors that influence the decisions regarding email personalisation are different.
Start by figuring out what the rate of returning visitors vs new visitors is for your store.
One of the best sources for this information is Google Analytics. Check what this rate for your store is. If returning visitor rate is greater than 30%, then you should be looking into adding personalisation. If the rate is less than 10%, you probably don’t need personalisation at this point.
An interesting case is when the rate of returning vs new visitors is between 10% and 30%.
If your store traffic data conforms to this pattern, you should be checking the revenue rate generated by the returning customers vs the new customers on your store. Frankly, you should be aware of what the value of this rate is, even outside of the “personalisation” context.
All major e-commerce platforms offer reports on the revenue rate for returning customer vs new customers. Shopify calls this first-time vs returning customer sales report.
It needs to be said, that over the years data has shown that returning customers are more valuable to an online shop than the new ones.
According to the 2015 report by Monetate, on average, returning customers drove 66% of revenue, while accounting for 48% of traffic. This makes a returning customer twice as valuable a new one, at least when it comes to generating revenue for the store. Again, on average.
Considering this, a good rule of thumb would be: one returning customer should be generating twice or more the revenue of a new customer. If this coefficient is lower for your online store, you should be looking into adding personalisation to extract more value out of your returning customers.
Use the formula below to calculate what’s the revenue coefficient for returning customers
Rrn is for your store.
Rr is revenue rate (in %) for returning customers,
Rn is revenue rate for new customers (in %),
Tr is traffic rate of returning customers (in %),
Tn is traffic rate of new customers (in %).
So, if I’m to take the data from the Monetate report, the
Rrn calculations become:
Don’t forget that averages are misleading, especially when online businesses of completely different profiles are included in such studies.
Don’t make decisions based on averages unless you know the distribution around the average is normal.
Nobody should be fooled by comparing apples to oranges.
So, a good practice for any online e-commerce business is to track their own sets of analytics measurements over time. Attempt to include this coefficient as a tracked indicator, as it will help you not only make the decision about personalisation but also it will help you understand your customer base better.
By the way, there are quite a few other conclusions that you as a shop owner can derive from this returning to new customers information. For example, this post by Optimove explains what a healthy balance of new vs existing customer revenue mix is.
So, to sum up:
The numbers I’ve used in the examples and calculations are at best, educated guesses based on our customer data. These should be taken with a grain of salt, and are obviously, not universally applicable. The data points for your store will be different, but the thought process will still stand.
I hope the information presented above has given you a few hints and data touch point to help you understand when and if an online e-commerce store should start adding personalisation as it continues to grow.
This concludes the four-part series about Product Recommendations widgets, their setup, sale attribution and a simple framework to decide about personalisation.
P.S.: If you happen to be a Shopify store owner, take a look at our Visely Product Recommendations app in the marketplace.