Challenge Overview
Challenge Objectives
-
Propose algorithm (model, training, testing) for predicting likely buyer company for a company on the market from a list of first round buyers
-
Propose handling of null values in test data
Project Background
Our client, a global investment bank, is looking to build a predictive analytics algorithm to obtain characteristic insight based on historic deal data taken from their CRM matching potential closing bidders on assets based on behavior in previous biddings.
-
Prediction algorithm will be used in client environment to predict most likely company to close the deal
-
Actual algorithm will be built in the next challenge
Technology Stack
-
Algorithm will be implemented using Python, R, or .NET stack - proposals for relevant libraries/frameworks are welcome but not mandatory
Individual requirements
Historic data is provided in two Excel sheets:
-
Bidder profiles - contains data about actual bids - Buyer’s lists from previous transactions, including details such as firm size, revenue, industry, etc. for each bidding firm and firm sold. Data also includes each bidding firm’s progress through each round until the final closing bid. This data is considered as ground truth. There is total of 42 variables (columns)
-
Seller profiles - contains data about firms sold in the bidding process - includes firm size, revenue, industry, etc. There is a total of 37 columns.
Main goal is to predict the most likely buyer, given the first round bids. Bidding process for one company is identified by the “ENGAGEMENT_NUMBER__C” column in the data set. First round bids are the ones with ROUND__C=”First”. Closed deals are marked with ROUND_C=”Closing”.
For example - a sale of company X (engagement number 80314) had total of 9 bids - 5 first round, 3 final round and one closing bid. Input for the prediction would be the 5 first round bids and the expected output is likelihood of each of the 5 companies to close the deal.
Here are some important properties of the companies available in the sample data:
-
Is the buyer a Financial (funds) or Strategic buyer
-
Attributes of financial buyers considered are:
-
Fund Size
-
Sector of Focus
-
Equity Check
-
- Usually 10-30% of fund size
-
- Also consider liquidity of assets
-
Relevant investment history and portfolio
-
Investment Style
-
- Growth vs value focus
-
- Defensive/distress investment strategy
-
Geography
-
- For financial buyers, distinction is usually between domestic and cross-border deals
-
-
Attributes of strategic buyers considered are:
-
Industry
-
- Specific work/service offering
-
Firm Size
-
Level of acquisitiveness
-
- Acquisition history
-
- Want aggressive acquisition rate, but recent large transaction might be hindrance
-
Financial Metrics
-
- Market Cap
-
- Revenue
-
- EBITDA
-
- EV multiples
-
Reason for acquisition
-
- Expansion
-
- Competition
-
- Augmentation
-
Geography
-
For strategic buyers, focus is usually on buyers within same region
-
-
Attributes of the seller company are:
-
Size of firm
-
Financial metrics
-
- Market Cap
-
- Revenue
-
- EBITDA
-
- EV multiples
-
Industry
-
Ownership
-
Relevant M&A history
-
Most important performance indicators are
- differences between financial and strategic buyers
- industry differences
- cross-border versus domestic deals
- acquisition/transaction history
Algorithm can use all the data available in the sample data set and you can also define secondary data features if that improves the algorithm. For example, you could introduce firm size ratio (bidder firm size / seller firm size) as a new measure (this is just an example, not a recommendation)
You should create a proposal for the algorithm that uses the historic data to predict the most likely buyers for a new company to be sold. Required algorithm properties are:
-
Handles null values gracefully - there are a lot of null values in some of the columns of ground truth data
-
Ranks the buyers on the list - predict which buyer is most likely to close the deal
-
Algorithm output should show why the is high/low (in terms of different factors considered)
You should submit a document with details for the proposed algorithm and/or a proof of concept solution, pseudo-code or any documentation that helps illustrate your proposal.
Your proposal will be evaluated based on the following criteria:
-
How much it uses the historic data - have you analyzed the provided data set and used gained insights from the data
-
Quality of the ranking system - is the ranking information in the output clear and does it explain why a buyer company has a high likelihood of closing the deal
-
Expected accuracy based on test data matches - a subjective measure - how likely is the algorithm output to match buyers in the sample data (if trained with a subset of sample data for example)
-
How detailed is documentation for the algorithm - it should have sufficient information for algorithm implementation in the next challenge