Basics of Recommendation Platform and its Challenges

Posted on Posted in Travel Recommendation

As Internet becomes more ubiquitous and online commerce becomes a norm, we as buyers end up leaving a trail of exciting data for businesses to exploit. This data (also called big data) has created a world of opportunities for smart businesses to understand their customers and thus

  • Target the right customer
  • Build more relevant products
  • Customise to the long tail

The other side of this analysis is now used by Big Data companies to:

  • Benchmark products
  • Build Recommendation Platforms
  • Construct predictive analysis based solutions

THE RECOMMENDATION PLATFORM:

In this blog lets focus on Recommendation Platforms and how they will form a core piece of the customers’ decisioning process in the days to come. Recommendation platforms will become increasingly important as products jostle for that limited ‘me time’ of their prospective buyers.

As online sales become more consultative (and less transactional), it becomes important for the products sold online to build strong customer endorsements. This happens when businesses encourage their customers to write and share their experiences of the product. By aggregating these experiences, Big Data companies like KePSLA are building smart & cognitive platforms that can help consumers to narrow a complex decision to just a few recommendations.

Big Data allows us to do recommendations with more precision like never seen before. Our key is to reduce Data to its smallest unit (and funnily we call it Small Data). The KePSLA recommendation platform suggests a few data points out of a large pool of data that is precisely relevant to the persona of the buyer almost similar to how Amazon recommends your next book out of its list of (approximately) 13,35,000 titles that it has at any time on its catalogue.

The KePSLA Travel recommendation Platform is content based and uses the power of collaborative filtering. While we have seen, these solutions used successfully by Media and the Retail Industry (of the likes of Netflix & Amazon), KePSLA is probably one of the first technology companies in the world to build a state-of-the-art recommendation solution focussing on the hospitality and travel industry however, the underlying architecture is Industry agnostic.

THE CHALLENGE:

Two of the key challenges that we faced when we build the products were what we call

  • Data Voids

The power of our platform is based on the volume of data that we are able to locate before we can make an intelligent recommendation to the consumers. India being a relatively new market for eCommerce, one of our challenge is to source relevant data. Our partnerships with large travel content businesses like TripAdvisor and HolidayIQ have largely helped us to circumvent this problem. To add to that we now work with over 2500+ hotels in India and help them generate their own customer reviews by providing them Mobile Apps as well as Email & SMS driven solutions

  • Data Boils

Many a times even when there is ample amount of data and the Algos are not well written, the recommendation platform engines will throw our obvious answers. For example when traveller looks for a beach holiday, Goa becomes an obvious first choice.

While our first version of recommendation platform has seen fair success, we are now pushing the pedal to explore deeper opportunities in this space by focussing on machine learning solutions and easy contextual search through “Experience Tags” like Watering-Hole, Hungry Belly or through “Aspects” like Bar, Food, Pool etc. to bring more precisions for our customers. Two of India’s largest travel brands now use the KePSLA recommendation platform to deliver better search results and thus improve their look to book ratios. We look forward to sharing more on this in weeks to come.

Interested in consultation to improve your Online presence and Reviews ? Contact us – http://kepsla.com/contact/ or Drop a mail to KePSLA at marketing@kepsla.com

Leave a Reply

Your email address will not be published. Required fields are marked *