Private Equity

JMI Implementation

• Conducted detailed research on company and valued the company using SOTP method and company comparable.
• Valued 37% of investor’s equity value at $65M and also identified the realized gain of $61M during 19 yr holding period.
• Assessed that the company has one of its best investment in portfolio with MoIC of 32.6x and IRR of 39.2%.
• Based on analysis, JMI team identified that the size of market remains a constraint for further growth potential and company had reached its maturity in their strategic frames.
• JMI recommended two options:

JMI Implementation

• Began the search by classifying 10,000+ companies.
• Acquired the data list from multiple sources such as Bloomberg, industry associations, and trade journals.
• Filtered and cleaned the list down to 500 companies based on size and profitability.
• As per client’s request, focused on 5 industry segments to understand the dynamics.
• Profiled 66 companies with a special focus on management and their affiliations, company financials, operations, and valuation ratios.

JMI Implementation

• Performed screening for public and private entities in the renewable energy space and compiled a list based on their offerings.
• Filtered and reviewed renewable energy companies with a detailed analysis and history of all private equity investments and acquisitions in the sector.
• Computed their current and forward trading multiples and arrived at a range of exit multiples.
• Provided the range of valuation multiples based on the analysis and helped identify KPI enabling premium valuation.

Solution

JMI’s solution comprised of the following approach:

• Operational level data: Extracted operational level data such as sales statistics, inventory levels, ordering/purchasing data, working capital metrics, customer acquisition costs, and other KPIs for high-level scrutiny of portfolio companies.
• Mapping with the current trends: Operational level data of portfolio companies was dissected further and benchmarked with the latest trends and developments to gain deeper operational insights.
• Creating value: This lead to the identification of a specific focus areas for each portfolio company along with data insights on revenue and costs. Insights across the portfolio companies further assisted the PE firm to implement best practice of one portfolio company to all others leading to improvement in the overall portfolio.

JMI Implementation

• Extracted the historical data and analyzed various data like shipment order type, from - to details etc. to design daily warehouse stock and inventory reporting tool for warehouse managers. This gives them a comprehensive overview of the level and location of stock, and shows recommendations on where to store each item based on customer demand.
• Analytics team also extracted three years of historical shipping data to predict the number of in - and outbound shipping orders per warehouse over daily, weekly and monthly horizons. It helped the company to manage their labor planning and maintain the right level of staffing to deal with customer demand on any given day.
• Observed that during the summer long holiday in US, there was a spike in demand for automobile parts and required a huge workforce for better supply chain management.
• Built an optimization model for its client which helped them to solve complex distribution network problems and provided optimal solutions within a given set of parameters. JMI team also observed that 8 of its distribution center had lower sales compared to average sales figures because of longer deliverable time.
• Provided insights to clients for better positioning of its warehouses to minimize delivery time and costs across entire network.

JMI Implementation

• JMI Collected sales and promotional data across stores and blended them to have a uniform item-by-store-by-day level of granularity.
• Standardized data is obtained by pre-processing data to de-trend, de-seasonalize, and remove holiday effects.
• Built forecast models for baseline and promotional estimations, using a combination of time series and a frequency-based approach.
• Modelled separately for regional Black Friday-type promotional events and identify the most selling and profitable categories.
• Used store-level daily sales distribution indices to break down weekly to daily sales.
• Identified similar products for new products by clustering for 10+ features (like sales, regional preference, complement, substitute, promoted group, etc.) and modelling based on them.
• Automated the end-to-end process and enabled calibration for daily or weekly updates.

JMI Implementation

• Collected large amount of customer data of its e-commerce portfolio companies from a range of data sources such as transactional/purchase data, demographic data, customer interactions etc.
• Used JMI proprietary similarity metrics with clustering techniques and decision tree to define top segments based on customer characters, geo-factors, market trends, macro-economic factors and others.
• Identified that two of its e-commerce companies have lower market penetration in top tier cities which has a high impact on its top line revenue.
• Used classification methods to score the degree of customer’s propensity for purchasing particular products of each company based on historical data obtained from the targeted market segments.
• Implemented time-series and artificial neural networks to predict revenue that is likely to result from newly identified segments and conducted back-testing using historical marketing data for validating the hypothesis.
• Cross-validated data and model accuracy using partitioned data and prebuilt ML/ AI models.

JMI Implementation

JMI’s solution comprised of the following three-phased approach:

• Data Extraction: Extracting data through web scrapping tool based on specific criteria such as company brand value, customer sentiment, product portfolio, social media posts, customer review, management team and other unstructured data.
• Predictive Analytics: By looking at relevant data patterns and trends, model can conduct early detection of the companies with high growth opportunities and decide whether it’s an opportunity worth pursuing or not.
• Deal Sourcing: JMI also helped the PE firm to identify the target companies & assisted in engaging company management.

JMI Implementation

JMI was involved in the end-to-end process of deal sourcing, including:

• A custom web scraping tool was built to extract data from multiple unstructured data sources based on specific criteria such as employee count, funding raised, investors, company description, company geographic location, and company sector.
• The tool filtered the results based on specific criteria such as company brand value, customer sentiment, product portfolio, social media posts, customer reviews, management team, and other unstructured data.
• By looking at relevant data patterns and trends, the model conducted early detection of companies with high growth opportunities and decided whether it was an opportunity worth pursuing or not.
• This was further refined through manual shortlisting by analysts based on investment criteria.
• Clustering analysis was also undertaken from the collected data points to further understand the geographic spread of companies shortlisted based on investment criteria. The analysis revealed that a lot of SaaS start-ups were getting set up in the states of Arizona, Texas, and Michigan.
• The shortlisted companies were added to the client's CRM portal, and custom outreach emails were sent to them.
• A deck of short notes for preparation before the client's call with the prospect company management was also shared with the client.
• Promising companies for the future were also flagged, labeled, and added to the CRM portal with an alarm to check on their status 6 months down the line.

JMI Implementation

• Prepared a detailed budget model to highlight the key revenue & cost drivers of business.
• Assisted in the identification of key earnings sensitivity parameters and KPIs.
• Created returns analysis for different types of scenarios.
• Constructed a flexible model with detailed drivers and switches, enabling evaluation of potential projects and investments.
• Conducted review of relevant indicators to ensure that the deal outlay value was quantified, and that capital spending met internal funding criteria.

JMI Implementation

• Assessed the client’s data preparation processes that relied on legacy PDF, spreadsheet, and CSV input formats. Used Bayesian frequency models to detect anomalies and discrepancies in input data.
• Acquired unstructured data from third-party sources like foot traffic, visitors, nearby amenities, and reviews that are likely to be associated with determining the price of the property.
• Analyzed historical data using statistical algorithms and ML methodologies to understand the rental yield, price appreciation, and occupancy rate.
• Developed a sentiment indicator tool and publicly available data (Zillow) and modeled property price and rental value on a one- to ten-year time scale.
• Developed an interactive web platform that delivers comparable analyses using a variety of different user inputs.

Solution

• Designed a solution on top of an applicant 360 data lake and replicated transactional data from the pool.
• Designed micro-batches to extract and curate features that were used to track the past performance of applicants and unique qualifiers such as the applicant’s location, past experiences, background, positivity index from answers provided, past recruitment status, etc.
• Core search engine models were trained on a regular interval (overnight batch), and trained artefacts were stored in the database. The training process used the hired status of the classifier to predict the likelihood of any candidate being shortlisted for any specific job.

• The JMI team also designed the "Search API" to take basic inputs from recruiters and retrieve the matching records. The service used AI runtime to shortlist and rank candidates further based on the probability of hiring within the context of search criteria.

Value Proposition

• The AI driven “search API “ was able to shortlist applicants with more than 90% accuracy. The solution was tested with historical data points using wide range of job profiles, funnels and applicants. The robust and scalable design of the solution outperformed the standard keyword and metadata-based search capabilities while offers low latency response.
• The solution minimized the time and effort of any recruiter to identify the ideal candidates for recruiting and window of hiring. It was primarily designed to support bulk recruitment initiatives where the recruiters need to qualify, appoint a large number of candidates for multiple positions and/ or locations within shortest possible timeline.
• The multistep AI based learning applied on validated history of applicants, their performance with interview and jobs offers deep insights on potential hires and extend first level of scrutiny as an intelligent hiring assistant. Such utilities boosted productivity significantly and minimized operational cost of the recruiting agencies.

The Challenges

• A financial company approached JMI with a research platform challenge. The application brought together market data with their own research and opinions into one view to create and evaluate their strategies.
• Their legacy implementation ran on several servers as a monolithic application, with a hand managed and complex infrastructure.
• The overall performance was poor, which is unacceptable in a trading research environment where timeliness, speed and accuracy are paramount. In addition, the ongoing cost of this setup was hard to justify.

JMI Implementation

• Conducted detailed fundamental research on company financials to get insights on revenue growth, cash flows and other key performance indicators.
• Identified that company had not only outpaced the industry growth by 10% in the past 5 years, but also had an above average operating margin of 24%.
• Found that the company’s next gen EUV product could extend the coverage of its customer roadmaps by 10-15 years which would help it to increase its market share.
• Identified that the company had been gaining market share for the past three years.
• Analyzed that the company has key risks from its limited number of customers. Any business failure or insolvency of one of their main customers could result in significant credit losses for the company.
• Researched the impact of current regulation on exports and found that the US government could require a license for many high-tech products shipped to China from other countries if US made components constitute more than 25% of the value.
• Based on research and management commentary, JMI found that company did not meet the 25% threshold but the US department might consider lowering the 25% threshold in some cases which could be a key risk in its future growth.
• Correlated the company’s enterprise value and EBITDA with competitors to understand its valuation with operating metrics.