Online shoppers love choice, but frequently face the paradox of choice — it can actually be harder to buy something when there are too many options. New AI solutions such as smarter search or personalization can improve the buyer’s journey immensely. These are two of the main topics that Algolia CTO, Sean Mullaney, discussed recently with Michael Krigsman on the CXOTalk podcast which is now available on YouTube or your preferred podcast app.
Topics covered by the speakers include:
Why search and discovery are important for ecommerce
The paradox of choice
How AI solutions such as ChatGPT has affected the industry
How natural language search improves product discovery
Traditional search vs AI search powered by vectors
Scaling vector search
Addressing the long tail challenge
Personalization: across multiple sessions and real-time personalization
Measuring ROI and KPIs for personalization and AI
Ethical and privacy considerations
The future of search
A brief synopsis of themes
Ecommerce businesses have wrestled with personalization and discovery for years. Companies such as Amazon, Netflix, and eBay were early adopters of AI and could afford to invest in the infrastructure and resources needed to capitalize on these technologies — Amazon launched its recommender platform back in 2003. However, the technology was out of reach for most businesses.
Twenty years later, everything has changed. Now, any company can leverage even more powerful technologies at a fraction of the cost. Whereas traditional search solutions could only match keywords exactly, new AI search understands intent and concepts. Shoppers will be able to input all kinds of queries — from symptoms to ideas – to get great results. For example, to a traditional search engine “milk chocolate” and “chocolate milk” are the same thing — just inverted words — but an AI search engine understands these are categorically different products.
Consumers are using longer, more complex search phrases today, in part driven by voice search technologies such as Google Home Dot or Amazon Alexa. Longer, more complex queries are sometimes described as the search long tail, which can represent as much as 70% of a retail website’s queries. AI with natural language processing and natural language understanding can work quite well now — even if searchers don’t use exact keywords.
Personalization also gets a boost with AI. Both search and personalization are able to leverage data faster than people. Both technologies improve over time. The queries, clicks, conversions, signups, and other event-driven data provides ongoing learning to improve results. Even single-session visitors can get personalized recommendations or personalized search results; using AI, the search and recommendation engine will know what products or queries are converting — so it can use that data to drive better results.
As companies adopt AI, they will want to rapidly experiment to determine which algorithms drive the best business outcomes. For example, Algolia has built-in A/B testing features to compare the outcomes of keyword-only vs AI search results, compare different AI algorithms, personalized vs non-personalized results, or any combination.
You can listen to the full podcast here: CXO podcast