Sajari's PosNegTracking aims to provide flexible tracking tokens which can be used in novel use-cases to provide both useful analytics and to feed Sajari's machine learning algorithm data about the performance of search results. The two primary features for advanced use-cases are multi-use and long lived tokens.
PosNeg tokens can be used more than once. This means that a single token can be used to send positive signals of increasing weight as a user moves through a funnel.
Additionally, tokens can be stored for long durations in order to attribute search queries with results even over long session times. For example, if a user performs a webstore search and finds a good result but doesn't actually purchase the product until much later.
This example includes three files of a shopping cart funnel where PosNegTokens are used to track search results (store products) as they move through the search -> add to cart -> purchase funnel. Additionally, localstorage is used to persist tokens long-term in case the initial search and final purchase occur over separate browser sessions.
On a search result screen, we can use the sajari pre-built support rendering search results which includes storing PosNeg tokens of clicked results into localstorage.
Then on a product details page, we do two things. First, we submit any pending positive tokens associated with result clicks from the previous screen. This submit lets the Sajari engine know that this was likely a good result for the search term. Then, when a customer adds a product to their cart, we submit the token with a higher weight. This gives a stronger signal for this result. The customer has moved closer to the end of the funnel.
Unlike other examples which render Sajari search results, this is purely an example of how your own brand's storefront might be setup.
Once a customer reaches the end point of the funnel, the cart checkout confirmation screen, we submit one more token with a greater weight indicating a successful purchase.
Unlike other examples which render Sajari search results, this is purely an example of how your own brand's checkout might be setup.