This post was written by Alex Ioannides, our Director of Data Science & Market Design
At Perfect Channel, our aim is to build and maintain optimal marketplaces for our clients. Our definition of an optimal marketplace is one where all buyers and sellers have confidence that they can always have access to ‘fair’ market prices whenever they need to trade.
There are many elements to achieving an optimal marketplace, such as a well-advised combination of:
- Marketplace Design – where we seek to configure the right transactional (or trading) mechanisms to produce optimal (fair) price discovery; and
- Marketplace Analytics – where we choose the right information for buyers and sellers to use for making the best-possible decisions surrounding price and inventory management.
Together with the most appropriate way of bringing these elements together, with a unified user experience that matches the natural workflows of participants in your market.
As we have stated before, optimisation of online marketplaces is far from easy. In addition to the right platform, you must have an effective business strategy, robust analytical technology and above all, patience. With all of these considerations we are writing a series of blogs – the first we are focusing on is the role of Marketplace Analytics.
A lot of people come to us asking for advice on what analytics they need or what we could produce using their data. Our approach is always to focus on working with our clients to understand the decision-making-processes (or workflows) of their clients – i.e. the buyers and sellers in their marketplace – so that we will know how and where to deliver data most effectively – i.e. metrics and predictions that will actually be used and become valuable to the marketplaces’ users.
This might be as simple as knowing what contextual market data buyers are mindful of when keying-in bids and making it available on the same page that bids are entered; or, it may involve building tools that use machine learning to predict likely future market scenarios, that sellers can use to assess the impact of trading today vs. trading next period, etc.
We divide Marketplace Analytics into three types:
- information that helps you understand what is happening in the market – e.g. performance metrics and price indices;
- information that assists with making the best future decisions – e.g. forecasting metrics and trade prices, etc; and,
- automation for optimising the performance of potential trades (or live listings) that are already ‘in the wild’ – e.g. listing and inventory recommendations, and automated buyer-to-seller matching algorithms.
We will be diving into these areas over the coming series of posts as we try and impart some of our understanding.
It should be noted, that we often refer to the use of ‘auction data’, but this should not be perceived as a restriction. At Perfect Channel we view all types of transactional (or trading) mechanism as a type of ‘auction’, whether it’s a bid for a fixed-price (buy-it-now), an English ascending auction, or for a supply-demand batch market. Some types of auction yield more preference information than others – and we integrate this whenever it is possible – but all of our approaches to marketplace analytics are designed to be agnostic to the chosen type(s) of trading mechanism deployed for your marketplace.
For more content: