Founding

ForecastMarket was founded by James Young, a data scientist with extensive background in applying machine learning in enterprise forecasting and scientific use cases. The idea is relatively simple:

• Hiring a consultant or buying software to make recurring forecasts can be expensive and they may not fairly compare their performance.
• The cost of a team is mostly for the time it takes to customize a pre-existing approach. 
• There can be massive cost savings for forecast users if these pre-existing models are deployed in a easily usable and scalable way.
• The forecast user benefits from these cost savings by making sure their data is in our standardized format. A 3 column csv file with id, Date, and value.
• Because of the standardized format of data input, our models don't have to be customized each time, saving time and money. We can also relatively quickly benchmark how different models perform on different forecast example problems to give users an idea what might perform best for them.

Despite the cost savings and transparency of the above scheme, there might be reasons for a company to keep forecasting "In House", especially at larger companies. Here are some of those reasons and how ForecastMarket can still help.

• Data Privacy:  You could use placeholders for the id's and map back to the original id's after forecasts are made. Additionally you could forecast existing models residuals, which obfuscates the true values of what you're forecasting but can still improve your forecasts by finding patterns in the way the existing model is wrong.
• Extreme ROI from Accuracy: In cases where every % of accuracy means a huge amount of money saved or earned, having a specialized team makes sense. ForecastMarket can be used to provide benchmark performance of our best models, against which you can demonstrate your improvement.

Short-Term Vision

As a strategic-decision maker, you need to gather data, predict outcomes, and plan for the future, but these tasks can be repetitive and error-prone without the help of data science and machine learning. Hiring an expensive data science team or consultant may not be feasible, but ForecastMarket offers a more affordable option. Through our user-friendly API marketplace, you can easily apply advanced solutions, including machine learning, through our portal and in some cases even excel or Google sheets. By comparing the performance, speed, and price of different solutions, you can find the best solution for your needs and get the benefits of an experienced data scientist at a fraction of the cost.

Long-Term Vision

The long-term vision is to expand to a multi-vendor marketplace, meaning multiple data scientists can sell their API service through this website. This will be mutually beneficial to both customers and service providers. Customers will have a wider variety of options and competing providers with visible performance metrics. This should drive prices down and lead to more accurate models. Service providers will benefit from having a standardized marketplace which provides a fair comparison of the benefits of different models in an unbiased way. Both customers and service providers will benefit from the speed-up of the matching process.