Time to find your inner cyborg…?
Analysis into how Hg and Two Six Capital are applying data science to better understand and improve business revenue drivers…
Private equity has long been a traditional, people-centric industry, but there are signs that the incredible advances in computing power and data science are being utilised to better understand company growth and revenue drivers. Just as the hedge fund industry has seen quantitative managers rise to prominence, led by Two Sigma and DE Shaw, bringing scientific rigour to public market investing, private equity is beginning to do the same.
One firm spearheading the data science approach in private equity investing is Hg, one of Europe’s leading technology investors.
Four years ago, Hg set up a dedicated Data Analytics team to help Hg's companies harness data and data science. This capability currently consists of a team of data scientists, data engineers off-shore, and Hg's cloud data platform and tooling.
Data science is a strategic enabler
Christopher Kindt is a Director in the Portfolio Value Creation team at Hg and leads Hg’s data team. In the last few years, there has been a spike in engagement as Hg companies start to harness the power of cloud data platforms. AI deployment, sales force effectiveness, pricing and new market entry are just some of the areas Hg’s data analytics are working on with company boards and senior management to drive value.
“We have a unique opportunity at Hg given that we are investors not just in software companies but more specifically in workflow software companies. Using cloud technology, we can now harness that incredibly rich seam of workflow data pulsing through our portfolio companies.”
The modus operandi at Hg is to harness the power of data science, which it views as an important strategic enabler for companies. In brief, machine learning, when applied to relevant use cases, can lead to tangible data-driven outcomes as companies seek to gain a competitive edge.
“Our mission is to harness the value potential of the vast amount of rich data following through our portfolio companies,” asserts Kindt.
“We have a unique opportunity at Hg given that we are investors not just in software companies but more specifically in workflow software companies. Using cloud technology, we have an incredibly rich seam of workflow data flowing through our portfolio companies with which to harness and make the most of.
“It is as much about sharpening our due diligence as it is about adding value, post-investment, to our portfolio companies. We are seeking to have impact with our portfolio companies within quarters, or ideally, months, as opposed to the years we were previously used to.”
Large-scale data manipulation
Over in San Francisco, Two Six Capital is pushing the boundaries of data science. Established in 2013, the firm uses a state-of-the-art platform designed to bring data science front and centre into the PE investment model, and move away from reliance on Microsoft Excel spreadsheets.
Two Six Capital’s cloud-enabled platform is supporting PE groups and, in helping to meld man with machine; providing what the firm likes to refer to as ‘cyborg-like’ sophistication.
Using 18 statistical machine learning models to understand data, Two Six is able to bring high definition insight to better understand company revenue drivers. It does this by ingesting 143 billion-plus data transaction points on the platform. So far, it has worked on over USD32 billion of completed PE transactions.
Sajjad Jaffer is a co-founder and managing partner of the advisory and investment firm. His inspiration for Two Six Capital originated during his time studying for an MBA at Wharton.
“Using our cloud platform we can ingest billions of rows of data for large-scale data manipulation. Our platform can slice and dice any permutation of revenue drivers. By counting transactions by individual customer, by individual product, by individual channel or store location, we are able to granularly see what the revenue and growth prospects are for a company,” explains Jaffer.
PE’s cyborg moment
Whereas humans are fantastic at asking the right questions and applying judgment to analyse a company, machines are purpose built to take billions of data points and run statistical analyses to reveal where the opportunities might be to unlock value.
Moreover, data science can improve the efficacy of how PE groups approach due diligence.
“Using the cyborg approach we are able to help PE managers pinpoint the winners and to walk away from the laggards when doing pre-deal due diligence. On the value creation side, post-investment, our clients can ask, ‘How can we best use this cyborg approach to turn the forecast valuation multiple from 2x to 5x?’
“In some situations, we’ve worked with a portfolio company for a year or more to drive transformation in the business, leading to EBITDA multiple improvement from the moment we did due diligence, through to value creation and vendor due diligence,” explains Jaffer.
Hg uses a data transformation leader that partners with each of its portfolio companies, to help connect the technical world data with the commercial day-to-day world of the sales and management team. This first stage involves shaping the company’s data strategy.
The next step, using Hg’s cloud data infrastructure template, is to come up with a technical blueprint, for the company to implement.
“Then we move into project execution mode to help the company kick start their data journey,” says Kindt. “Depending on what the strategy, we assemble a team of data scientists, data engineers and visualisation experts to work through a programme of use cases.
“The use cases we develop are about driving EBITDA impact. For some portfolios companies we’ve worked with for a longer period, we can, with confidence, point to at least 10 per cent of their EBITDA growth coming from the data projects we’ve initiated. It is a tangible value driver.”
Finding eureka insights
Jaffer says that in recent weeks, Two Six has seen a spike in interest among PE groups keen to deploy data science to portfolio construction and management. This is understandable given the havoc wrought by Covid-19. Understanding operating revenues is vital.
Just today, for example, CNN reported that SoftBank expects to lose USD6.6 billion on its ill-fated investment in WeWork; the office sharing group that has been hit hard by coronavirus.
Jaffer says the questions PE groups are now asking revolve around value creation in PE portfolios; from pre-deal due diligence to vendor due diligence and everything in between.
“Right now the focus is on this ‘everything in between’ so that PE managers can be more mindful of how to position companies for sale,” he says.
In a world where revenue has fallen precipitously, one critical question PE investors have is, ‘What is the right cost structure to sustain a company in the current environment?’ The data can help portfolio companies with cost containment scenario planning, which, as Jaffer states, “is really important right now”.
“A second key question is, ‘What does the recovery look like and how do I use big data to understand what this business could look like, and position it for the next wave of customer demand?’ adds Jaffer.
In an article posted by Wharton last July, Jaffer compared the continual monitoring of companies to trying to lose weight; “If you really want to see results, you need to step on the scale every day,” he was quoted as saying.
Applying data analytics to uncover a Eureka moment will vary, depending on how PE groups choose to develop use cases for their operating companies.
Those Eureka moments might relate to growth; ie Where is this company on the saturation curve for each of the market segments it serves? How much headroom growth is there?
They might also relate to customer quality; What is the true retention rate, or the true churn of performance?
“Especially when looking at subscription or SaaS-based companies, a few basis points in churn estimation can lead to a significant valuation gap,” says Jaffer.
One example of how Hg applies data analytics would be using machine learning techniques for cross-sales propensity modelling.
“One company had a good sales team and a competent sales leader who was somewhat skeptical about our machine learning models. But since we implemented the project we have seen an 83 per cent uplift in cross-selling compared to the previous baseline,” states Kindt.
There is still a huge amount of potential yet to be tapped in to as PE groups apply data science techniques to better understand their portfolio assets.
Looking ahead, gaining clarity on revenue and growth drivers could bring man and machine closer together, as PE groups tap into their inner cyborg.
“This presents an opportunity for us to think of ourselves as not just as PE company only 30 disparate companies but as a data centre of excellence, working with local teams and realizing cross-company opportunities… ie finding new customers, providing benchmarking insights.
“That is something we actively working on right now,” concludes Kindt.