Automating the matching process with data-driven comparisons and a simple wizard
With their Minimum Viable Product (MVP), our client had covered off the most basic requirement: get to market and establish the need for an online marketplace to assist users seeking specialists.
The functionality as designed was working well, and favourable matches were being made, but it was reliant on on user inputs and missing granularity as a result of reliance upon hard-coded filter fields.
It was also evident that the platform was punching below its weight - with integration to data directly out of vendor CRMs, there was no need to serve users solely via filter searches.
Our client was now looking to automate the matching and ranking processes to best serve their growing client base.
Depending on your particular property details, our matching algorithm can be very specific about who will achieve the best result, even within the same office - Helmsmen
Our client undertook research into inspirational data-driven comparison ranking sites from several market sectors before organising a set of questions to be posed by a wizard and deciding upon their matching ‘rules’.
We completed a research and design stage with our business analyst (BA) and organised all of the client requirements into thirty or so tickets under a single ‘epic’. Technical solutions were found, exceptions or issues anticipated and complexity estimated as we went.
The R&D, BA and build out phases were completed collaboratively over the course of four two-week sprints with automated deployments to a test environment and direct commentary from our client in our agile management tool, JIRA, (as well as plenty of fruitful conversations).
The wizard made the search process as easy as a few clicks for the user while the matching algorithm paved the way for future iterations based on more complex algorithms run against much larger data sets adapted expressly for the ranking process.
With plenty of data already being collated within the application, there was a clear capacity for the platform to move towards providing data-driven comparisons rather than filter-led match results.
Automating that process via a user interface type known as a wizard with a matching algorithm driving the resulting comparison rankings was the most sensible step to take in order to assist the user.
A set of matching ‘rules’ were codified so as to search for, compare and rank results for users based on their chosen wizard inputs. Best-fit suggestions would display at the end of the process, collected data would be sent to our client, and the user would be given the option to connect with their favourite specialist.
Re-imagining the agent-seller relationship with proptech