The shorter investment lifecycle of private-market investments has made it necessary for participants to access analytics and other data-led processes at speed.
The obvious focus in achieving that has been on developing artificial intelligence applications. But piloting initiatives on evolving models can take time. Organisations want to test their applications to know they will work without error before rolling them out into production.
In large part, pilots will be used to ensure the data that is fed into the models is trusted. Without that, organisations can’t be sure that their applications will generate accurate outcomes and decision-making based on them will be erroneous.
But some organisations are already using applications they have developed without extended testing. Not content with waiting for “perfect data” to arrive to validate the efficacy of their initiatives, these companies have seized on models and methodologies that have already proved their trustworthiness as they have developed new data products, said Siva Ilango, partner at JMAN Group, which services the data needs of private market participants.“It doesn’t mean that you need to wait for two years to get to a foundation to prove the commercial value of your product,” Ilango told Data Management Insight.
Traditional Statistical Methods
AI has developed so quickly that recent generative AI (GenAI) models and established agentic applications can be deployed, along with traditional statistical data science methods, to solve for the “low-hanging fruit” of private-market data challenges.
Advances already made in AI means that an iterative approach can be taken to a new data product; it can be rolled out incrementally to a handful of non-essential users and gradually expanded and refined as its performance is assessed. In the same way, an application can be developed to work with a confined set of data before the whole gamut of data feeds are onboarded.By prioritising early value over achieving 100 per cent perfection in engineering and data, products can be brought to market quicker.
This use case-led approach is one that JMAN has brought to some of its mid-market customers. For instance, Ilango explained that for some funds the company has introduced data applications to the financial teams before extending them out to the broader organisation.
“If you wait for the data or application to be perfect, you are losing opportunities,” said Ilango.
The iterative approach is also a pragmatic solution to accommodating innovations to the technology built into those applications.
“If you’re looking for perfectionism you have to be aware that if you are building it today it may change in six months. It may change in 12 months,” he said. “Large corporates have the luxury to do that.
“But especially in the PE space, you want something to go quickly. The valuation period of an investment is three to five years. So if you’re spending two years building the foundation, you’re already gone.”
Portfolio Companies
The strategy has been successfully applied to business-to-business operations when the regulatory scrutiny has been less stringent. This is applicable to many portfolio companies, who will lack the sort of integrated data setups that GPs will expect of them.
It’s equally applicable to some parts of the private markets ecosystem, especially in the back-office operations that rely on customer relationship management platforms, said Ilango. In that context, iterated applications can be introduced because the data is not updated as quickly as in the front office.
“What we see mostly when we speak with the portfolio companies is that they are all doing industry migration,” he said. “Then the fund comes in saying, ‘okay, we’re trying to capture that, can we build the data and the technology on top of it?’”
Subscribe to our newsletter
