3 Examples of E-Commerce Personalization
In our work at Datacop we often come to discuss personalization with our clients. We see that many in the e-commerce industry are aware of the importance of personalization. A survey of marketing professionals has found that 88% of marketers said that they perceived a measurable lift through personalization in business results and 61% also said that personalization helped them deliver a better customer experience overall. (source)
On the other side, customers expect personalization. According to an Accenture Customer survey, 91% of online customers prefer to purchase from brands that offer a personalized experience. (source) Personalization promotes loyalty to a brand, 44% of surveyed customers they would likely become repeated customers if their experience was personalized. Additionally, surveys suggest that personalization increases the amount a customer is willing to spend more with a brand. In a survey, 40% of customers said that they purchased something more expensive from an e-shop because of a personalized experience. (source) The digital giants such as Amazon, Facebook and Google leverage data collected about their customers for personalization of their Shopping Recommendations and News Feeds to great effect on user's adoption of their respective platforms. In doing so, they have raised customer's expectations of what is possible.
Even though personalization has become a popular talking point of many in the industry, still many professionals do not fully understand personalization well. Many are not aware about the technical limitations of personalization. Additionally, the application of personalization into practice is difficult. Consider the many emails you have received where the only thing that was personalized about it was the name at the beginning. While that email is using personalisation, it does not necessarily mean it is effective. Think of the many IP-based personalisation on the web that let you know where you are connecting from, but do not really work with that information to provide you local recommendations. It is true that now the website is personalized based on geography - but how does that connect to the needs of the customer on your site? This is what we refer to as “vanity personalization”, where the content is personalized without adding any added benefit to the customer - are often ineffective at driving business results.
This post will aim to clarify personalization for ecommerce teams. In the post we will discuss
What is Personalization?
Two Personalization Use-Cases in Fashion E-Commerce
One Personalization Use-Case in Childcare E-Commerce
What is Personalization?
Personalization, once limited to targeted offers, now extends to the entire customer experience - throughout their interactions with a brand with multiple touchpoints. In the best personalised experiences, retailers make the customer part of the dialogue and leverage data to create one-to-one personalization. Customers receive offers that are targeted not just at customers like them, with brands targeting at the segment level with broad-based offers, but at them as individuals, with products, offers, and communications that are uniquely relevant to them. Although, in practice both segment-level personalization and 1:1 personalization have proven to be effective.
Where did it originate from? Personalization as a concept has been around longer than the internet. It used to primarily mean targeted offers delivered through post-mailed catalogues or in-store promotions. Marketers and Operational Researchers of the analogue era, working in big-chain retailers like Costco in the US or Tesco in the UK pioneered first large-scale automated "personalization" as we understand it today in the digital era. Personalized (mostly a segment-level- targeted) offer/coupon personalization was enabled by a clever innovation in data technology - the retail customer club cards. What these clubcards did, was that they incentivized customers to scan their card at their checkout at the store - as they would benefit from small but exclusive rewards and discounts on everyday items. This enabled the retailer to create a CRM for its clubcard members with the customer's clubcard ID as the customer ID - connecting their transactions across time and space (different branches) and its attributes (items, prices, categories, locations of purchase, etc.) that could be analyzed to start segmenting customers into meaningful buckets based on their demographic and purchase data (not-a-family, family, has a pet, has a garden, etc.) and in turn then design offers tailor-made for those customer segments.
Google Ngram Trend for the words 'personalisation' (blue) & 'ecommerce' (red) between the years 1950 - 2019. In this chart we can see that personalisation preceded e-commerce as a concept and also that the growth of e-commerce, correlated strongly with the growth of personalisation. Google Ngram is a search engine that charts word frequencies from a large corpus of books that were printed between 1500 and 2019.
Customer Data & Customer Experience Platforms
Today, e-commerce firms use customer data platforms and customer experience platforms to enable personalization at scale and provide a consistent personalization experience across most or all marketing channels. A Customer Data Platform allowing the brand to collect and retain information on their customers over time and unify the various data sources customers generate about their behaviour into a single customer record. The ID for a single customer in the case of e-commerce is usually the customers email address. The adoption of CDP software by the e-commerce industry in the 2010s was a similar "aha" moment as when Tesco introduced the Tesco Clubcard. The clubcard's ID is "replaced" in most e-shops by the customer's email address. The automated collection of customer data across time, at scale and in real time enables digital firms to create a sort of an "artificial memory that never forgets" (unless asked so by the data subject). When this data flow is then integrated with an experience platform that marketing channel like a Website, the content of the website will request information from the CDP to know what to display. The collection of the data and its integration with the channels and touchpoints customers face are the two important technical challenges these software overcome for e-commerce firms.
What does good personalization mean in practice?
We talked about what personalization is and where it came from, yet still it may feel a bit abstract. One reason for this is that personalization looks different from e-shop to e-shop, depending on the type of products/services they are selling, their price point, how many products they have, etc. It depends on the data the e-shop can collects, the needs of their customer segments and the channels the e-shop uses to communicate with their customers. At Datacop we believe that effective personalization, as opposed to vanity personalization, needs to fulfill the following two criteria:
a) meaningfully differentiate content or the timing of the content to different segments by leveraging customer data.
b) improve the value of the experience or communication for the customer.
Lastly, because personalization requires time and skill to implement, effective personalization must justify its investment. Therefore, whenever possible we encourage our clients to split A/B test their personalization, to understand its impact on the customer journey and the bottom line. Now with all of this theory out of the way, let's have a look at three examples of what effective personalization looks like in practice:
Personalizing Emails in Fashion E-Commerce
Many fashion e-shop offer an exclusive discount coupon to first-time purchasers in their Cart Abandonment automated email scenario. The discount is usually between 5-15% depending on the discount strategy of the brand. This tactic improves the conversion rate of the customers who abandoned their cart. The first example presented in this article is the personalization of the subject line of this kind of automated cart abandonment email with a discount, by automatically calculating the value of the discount based on the last total price in the customer's basket. Now instead of the Subject Line reading "10% off your pending order" it would read "(( value of the customer's basket * 0.1 )) off your pending order".
This personalization was split tested 50/50, half of the customers who abandoned their cart would receive the original email with the 10% discount and the other half would receive the second email with the same discount except the concrete value of the discount would be calculated and displayed to each customer based on their last cart_update event. The half that received the personalized version of the email had twice as many orders as the one with a non-personalized blanket 10% off. The A/B test was turned on long enough to observe statistical significance on the difference between the cohorts. Even a small but meaningful change in an already well functioning marketing automation can bring in interesting value.
This first example fits both of our Datacop criteria of effective personalization:
a) 1:1 personalization; each email sent will be different based on the value of the cart the customer abandoned
b) the automatically recalculated value communicates more information with the same space
Personalization of the Subject Line can be surprisingly powerful, as many decide whether they will interact with an unexpected email or not based on their initial impression of the content.
It is common for fashion e-shops that sell shoes to have an end-of-inventory sale. This is because shoe sales often follow a seasonal pattern and there are almost always certain sizes left over from most styles by the end of their season. These inventory sales are often promoted by email to existing customers or potential customers who have signed up the brand's newsletter. One common issue of these emails is that, if a customer is enticed by a shoe style they liked in the email or on the landing page of the website; quite often in these inventory sales they will not find the shoe they like in the size that the shop currently offers in the sale. This often leads customers to disappointment and abandonment of the site.
The second example in this article is the personalization of this kind of end-of-inventory sale email and landing page, by the customer's shoe size - data that can be leveraged from their last transaction. Now instead of the subject line and the visual content of the email being the same across the entire email base - "Up to 60% off" those with a purchase will see "Up to 60% off your (( size of the customer's last purchase_item )) size. Additionally, the landing pages of these personalized emails will already have their respective size filter turned on, thus the customer will only see items that are actually available to them from the sale, saving them from potential disappointment if they took a liking in one of the items.
This email campaign was split tested 50/50, half of the customers whose size was identifiable saw the same thing as the customers with no data and the other half saw the personalized version of the campaign. The personalized version of the email had a 50% increase in orders and an approximate doubling of the Open Rate% and Click Rate% in comparison to the control group. The reason this works, is that the personalization imitates the personal touch of a shopping assistant in a good shoe store. A good shopping assistant acts in a similar way to a CDP, they would remember when a customer has visited the store for the first time or whether they are returning and if the customer is returning an excellent shopping assistant would remember their preferences and help them ease their selection process to a best fit with the customer's needs.
This second example fits both of our Datacop criteria of effective personalization:
a) segment based personalization, each email sent will be different based on the size of the last shoe purchased with the brand
b) the size based personalization saves customers approximately 2 friction clicks in their journey, saves potential disappointment from finding a product they would want, only for it to be unavailable in their size.
Personalizing in Childcare E-Commerce
In childcare e-commerce an advanced way to personalize the shopping experience and email experience for customers is by requesting the date of birth of their child and storing that value with the customer record of their parents, the true customer. The first 18 years of our lives have a lot of predictable situation that our parents have to deal with. Childcare brands can derive the age of their customers' child or children from their date of birth. This kind of personalization differs from merely just effective personalization in that it utilizes data that the customer share with the brand willingly outside of the basic data collected through the act of using the brand's service or buying the brand's products. This data can be very helpful for a variety of purposes.
The age of the child can be effectively used in email personalization. Leading brands that cater to children at the baby stage of a child - i.e. before and after their birth have designed an automated email program that informs subscribed parents about common tips to think about depending on how many weeks they are from birth or after birth. At the same time these email offer a great opportunity for the brand to recommend relevant products to parents, depending on how many weeks from birth or after birth of their child they are. This has proven particularly successful with parents who have their first child. This kind of personalization isn't just about recommended products or greeting them with their name, but it actually offers a service - the kind an experienced sales assistant would have in a childcare shop for babies. These emails had high engagement rate in comparison to the brand's other automated emails. In addition these customers were more likely to purchase than customers who were not participating in the programme. However it was not tested. In this case, it is plausible that customers who were willing to share their baby's date of birth in the first place were much more likely to purchase anyway. At the same time it is known that people who make a small commitment first (sharing data) to a brand are more likely to make a bigger commitment down the line (making a purchase).
The age of child can be effectively used for analytical insights by studying e-shop behavior of parents that are in those different stages before and after birth. Because each event is stored with a timestamp in the same record as their children's date of birth, it is possible to calculate a time difference from birthdate for each event and thus infer what parents are looking for at each stage of their journey. This allows childcare brands to answer very precise questions about their customers: What are parents looking at a few weeks before birth? Does that differ from what they browse 40 weeks before birth? One of the findings we found was that parents two weeks before birth had three times the amount of Foods and Clothing viewed and added than parents in the 20-40 weeks before birth segment. Full evaluation of this kind of data could yield tens of interesting insights especially if it is paired with the existing qualitative knowledge in the brand regarding the needs of parents during their journey of 9 months before birth and 2 years after birth.
This third example fits both of our Datacop criteria of effective personalization:
a) segment based personalization, web and email experiences are modified depending on which stage the parents are in.
b) in this case the personalization is educational and reassuring to the parents who are going through a life changing journey.
An example of how a brand can capture data about their children. This web form asks for the name, day, month, year and gender of their child.
If you found this post useful
Subscribe and share our substack article with your colleagues in the industry :) !