Evolving from “one size fits all”
The web has historically been a centralized model for businesses to advertise and sell – one online destination for many to come to access content. Visit any of your favorite websites and you would see the same content as another user clear across the state or country. This standardization ensured a consistent experience whether a user is visiting to check sports scores or the latest performance of the stock market.
As the web evolved, so too did the user experience as this “one size fits all” approach was no longer the optimal strategy. From Facebook’s News Feed to YouTube & Netflix recommending videos to watch next based on your prior view history, the homogeneous content strategy began to change. Other businesses began to advance as well terms of how they matched users to the products & experiences they desired, leading to an improved experience.
A more personal approach
At Algolia, we have been focused on personalization for a while now, constantly thinking about how we bring customized search opportunities to our customers. Today, we’re excited to release a brand new version of Personalization, which lets customers use a list of inputs about a user’s behavior to influence the content that this person is going to see. More specifically, we now allow businesses to take a wide range of individual online behavior & preferences, and optimize the search ranking strategy accordingly to show more relevant results. This is accomplished by using a list of inputs about a user’s online behavior to influence the content that this individual will see. While there are many different types of general digital personalization tools (for dynamic website content, email layout etc.) our Algolia solution specifically focuses on search and discovery.
How to personalize
So how, and more importantly what can be personalized? The two most prominent types are personalized search and personalized lists. Personalized search applies to changing the search results to match the profile of a user once a search is performed. A personalized list by contrast displays a list of objects that match the profile of a user, but are outside of the search (think the “Recommended for You” choices Netflix provides upon entering your user account). While there are a number of inputs to potentially track, we are starting with the following initial events as potential inputs to personalize someone’s experience:
- Views: If a user has seen a product page, or a landing page
- Clicks: What a user has clicked on
- Conversion: When a user adds an item to its cart, makes a purchase, watches a video, reads an article, etc.
It is also worth mentioning that there are two big varieties of personalization strategies, that may or may not be combined together: reinforcement & discovery. Reinforcement relates to pushing items that you have directly interacted with individually (i.e you bookmarked a product) or as a group (you visited a landing page with multiple products on them) to try and increase the accuracy of results you are seeing.
Discovery by contrast is focused on highlighting items that are related to the items a user has directly interacted with. This works especially well for items that are paired together (need a case to compliment the new cell phone you just bought?) and focuses on discovering interesting items you might otherwise not have seen. All of these specifics are designed to let our users better control how the ranking will be impacted, using a mathematical approach to let businesses prioritize accordingly. While there are many other variables to track, we will look to potentially add more through future releases and iterations of the product.
One of the key benefits to our new Personalization release is how easy it is to configure and get up and running quickly. Most personalization strategies take a lot of work and development resources to get set up, and once live, it becomes even more burdensome to gain visibility and insight into your relevance strategy. With Algolia, all you need to do is set your event weights (the relative importance of each) and start to send us the events, and we do the rest. Best of all, once you’re live, you can iterate on your formula and fine tune it to your needs in an instant through our dashboard.
Customers using Personalization
Videdressing has seen value in getting a personalized experience up and running quickly, as Product Manager David Piry notes “One of the great things about Personalization with Algolia is how easy it is to get set-up and running – there’s not a ton of development work, just using data you are already collecting.” Personalization solves the issue of a “one size fits all approach” and increases relevancy by displaying content and objects your users are more likely to be interested in.
Want to learn more? Click here for details.