What is personalization?
Personalization is essentially used to refer to a variety of algorithms and strategies that web-based businesses use to customize the experience on their websites for different segments of the audience. For instance, an online fashion store might want to automatically show the men’s collection if it detects that a male user has logged on to the website. Another example would be an OTT platform that wants to recommend content to watch based on your location and past preferences. Yet another example of personalization would be a website that automatically displays the homepage in the national language of your country.
Personalization efforts usually help businesses improve conversion rates on their websites, ease the search for an appropriate product or improve the visitor experience of the platform. Personalization efforts often vary from personalizing coarsely for a group of audience segments or personalizing one-to-one based on various user attributes.
There are different kinds of personalization algorithms available in the market and one needs to clearly understand if personalization is the right strategy to grow their business.
How does personalization drive up conversion rates?
To understand how a website can benefit from personalization, consider the following example. Suppose, you have the option to build your homepage either in English or Spanish. Also, suppose that you have two audience segments. Half of your traffic comes from the USA which only understands the English version of the website and has an average conversion rate of 10%. The other half of your traffic comes from Spain which only understands the Spanish language and has an average conversion rate of 8%.
If you were to choose between the English and the Spanish version, you would probably go for the English one since users from the USA have a higher conversion rate as compared to users from Spain. Your net conversion rate from the English version would be 5% (since 50% of the Spanish user traffic is completely ditched), and from the Spanish version, it would be 4% (vice-versa).
However, if you were to personalize based on the country the visitor is coming from, you could cater to both the Spanish and the English visitors and get a combined conversion rate of 9%. This way, you’re not only offering an improved experience to different segments of your audience, but you’re also going for higher conversion rates.
What are the types of personalization algorithms available in the market?
- Rule-based personalization: A rule-based personalization engine takes as input the different variations of a webpage and a rule mapping to know which variation to display to which visitor. Such rules can be created using any attributes that are tracked over all visitors. As soon as a visitor is identified to belong to a particular segment, the personalization engine shows the corresponding variation to the visitor.
- AI personalization: An AI personalization engine takes as input the different variations of a webpage to show to different visitors and various visitor attributes. It then tries to learn which visitor attributes convert better on which variations. Hence, in the English-Spanish example (given above), an AI personalization would automatically learn that when the country is the USA, show the English version and when the country is Spain show the Spanish version. AI personalization usually works through multiple attributes and learns in real-time what would be the best version to show to your visitors.
- Recommendation engines: A recommendation engine is usually valuable for websites with a large inventory of different products or contents. Recommendation engines usually use visitor attributes to learn similarities between different visitors. When one visitor ends up liking a particular product/content, the recommendation engine learns and shows the same product/content to similar visitors. Other types of recommendation engines also utilize product-to-product similarity. So, if a visitor likes a particular product, similar or complementary products are recommended to the visitor.
- Survey-based personalization: A survey-based personalization engine often asks some questions to visitors before personalizing the page for them. For instance, an eCommerce website might ask visitors a couple of questions regarding what they are searching for. Based on the answers given by visitors, the website can automatically personalize the homepage so that visitors can conveniently find the product of their preference.
How are personalization and experimentation connected?
While personalization efforts can be focussed towards generally improving the visitor experience on the website, usually personalization efforts are focussed on optimizing a KPI metric which can be anything like conversion rates or revenue on a page. The broader culture of experimenting with novel product ideas to improve conversion rates is called experimentation and has been popularized in the market with different names such as A/B Testing and Conversion Rate Optimization. In this regard, personalization is also a product of advanced experimentation, and companies that have achieved some degree of success with vanilla A/B tests move towards personalization to further optimize their KPI metrics.
Usually, this starts with companies trying to analyze their A/B testing data over different segments of the audience. Even on VWO, you can filter your A/B test data based on different visitor attributes and observe how the conversion rates across segments differ from overall averages. In such analysis, if you find interesting patterns where particular segments behave differently from the average, you can start developing hypotheses towards how to cater to these particular segments better by showing them a different page (than what is shown to the general audience).
With VWO Personalize, you can build and trigger relevant, contextual experiences for your website visitors at the right place and at the right time. To understand more about this, take a free trial or request a demo with our product experts.