The Buy vs. Build Debate for Investment Software

One of the most common questions investment firms have around software is whether they should build tools in-house or buy externally. In this article, we’ll detail a few successful and unsuccessful internal tools we’ve seen, what it takes to build something in-house, and what is coming in software-for-investment land.

The basis for this article is discussions with over 730 investment firms across venture, growth equity, buyout, and publics about their tech stack. We’ve seen over 50 internal tools across the investment landscape – everything from basic model wrappers to Airtable replicas to advanced firm-wide operating systems. 

In our time in private equity, we’ve also been customers of every single tool out there, whether it be CRM (i.e. Dealcloud, Affinity), alternative data (Similarweb, AppAnnie, Consumer Edge), sourcing (i.e. Synaptic, Pitchbook), AI tooling (i.e. Hebbia).

On a macro level, here is the order of technology sophistication we’ve seen in the financial services category:

¹Only reflects a subset of large cap PE. There is generally a large divergence in terms of tech-savviness across large cap and upper middle-market private equity. There are many firms of similar scale to those listed above who are not sophisticated in terms of technology adoption / sophistication. This includes leading large cap firms such as CD&R, H&F and KKR.

‍Note: This analysis / thought piece does not include multi-managers, because they are ultimately of a different level of sophistication and budgeting (i.e. tens of millions of dollars annually). The goal of value-added software in publics to attain either i) information that others do not have access to or ii) information before others get it. This is ultimately a different game than the rest of investing / advisory work.‍

Examples of Strong Internal Tooling

Let us start with a few aspirational examples in which internal tooling is superior to all platforms available by existing software providers, and built custom to firm requirements.

1. Coatue’s Mosaic
   • De-facto operating system of Coatue – CRM / deal-tracking, alternative + public data aggregator, metrics benchmarking, HR system, portfolio monitoring, and overall system of record for the firm
   • The tool helps investment professionals to easily search, track, and outreach to companies, view broader Coatue portfolio performance, and benchmark public and private company data / metrics with ease 
   • We estimate the cost of data alone must be in the LSD millions alone annually across macro, social media, utilities, app, web traffic, reviews, among other types of bespoke data
   • On top of the data, we estimate Coatue has spent in the high tens of millions to build this tool (over a decade), and several million dollars each year to maintain across their on-site staff

2. Blackstone’s internal deal modeling tool
   • Allows users to instantly generate LBO and BX-style outputs and sensitivities using consensus estimates, with the ability to flex inputs such as growth rate, debt assumptions, share price, etc.
   • Allows investment team to instantly see if a public company’s share price appears undervalued based off preliminary view on consensus estimates, and view alongside semi-private alternative data
   • Anecdotally, frequently used by Blackstone MDs to see what consensus numbers imply in terms of valuation

3. McKinsey’s Lilli
   • Chat-based tool for internal McKinsey outputs – more advanced than other chat tools we’ve seen as it has:
     1. Ability to output directly into PPT, excel and word
     2. Direct references to relevant past McKinsey work and ability to see who had worked on those projects
     3. Extensive prompt library for common consultant prompts
   • Allows consultants (usage primarily comes at BA level) to quickly output a first cut of various outputs in McKinsey format (i.e. SoW, implementation plans, etc.) and immediately connect with other McKinsey employees who have put together relevant materials
   • We estimate at least HSD millions to low tens of millions cost to build

4. TCV’s DIG tool
   • One of the few internal CRMs / data aggregation platforms we’ve seen that is quite thoughtfully built
   • On top of standard CRM capabilities, also has Affinity like notes tracking/ interactions information as well as highly enriched alternative data (i.e. credit card data, mobile app data, people data, web data, open-source data, etc.)
   • Generates business score, which is a percentile ranked likelihood of investment based off predefined metrics (i.e. headcount growth, revenue growth, scale, metrics, etc.)
   • Allows investment professionals to quickly produce cuts on private data, track institutional knowledge, and get 80/20 of company performance

Examples of Unsuccessful In-House Building

On the other hand, we’ve seen many instances in which internal tooling becomes a major time and budget drain but also yields low adoption among investment teams.

There are many more than we can count, but here are a few examples.

1. Public Large Cap Private Equity Firm A
   • Internal chat-based tool that is effectively a white-labeled, secure version of ChatGPT (i.e. leveraging open-source model such as Llama rather than closed-source)
   • Next-to-no usage from investment professionals today
   • While ChatGPT is blocked at this firm, many investment professionals resort to Claude instead
   • Estimated cost is >$500K to build and >$300K to maintain.

2. Large Cap Private Equity Firm B
   • Deal sourcing tool that leverages data science to ostensibly identify companies that fit their investment criteria across growth equity and buyout.
   • Great marketing tool to LPs, but minimal usage from investment team
   • Estimated cost is $2M+ to create, and over $600K / year to maintain

3. Large Cap Private Equity Firm C
   • Built data-science / engineering team to work on various internal projects
   • 80% of teams were shuttered in ’22 for sensitive reasons
   • Estimated LSD millions budget drained

4. Leading Venture firm A
   • Built in-house CRM tool like Affinity with plugin into datasets
   • Ultimately, UI was clunky and quite laggy, so the firm ended up switching back to Affinity.
   • Estimated $500K burned over 2 years and many investments professional hours spent building

5. Leading Publicly Traded Investment Bank A
   • ‍Company searching tool that allows users to semantically search across using keywords (i.e. industry, scale) and presents similar companies based off data and tagging from Pitchbook and Factset
   • While core intent and idea for this tool is interesting, end product is clunky with poor UI, and therefore has very low engagement / adoption

‍Note: Given the sensitive nature of this topic, all firm names here are redacted.

While each of these tools had different reasons for failure, here are the most common reason for failure we’ve seen:

1. Offshoring work to low-cost regions – quality of work will naturally be lower, and engineering is not like finance in which a good analyst is 40% better than an average one, the delta scales exponentially
2. Insufficient investment professional oversight – dedicated investment professional involvement is an absolute prerequisite to build something with adoption
3. Lack of budget – anything less than $1M annually of budget is insufficient (unless you are a venture firm building a basic sourcing tool / CRM)
4. Unintuitive UI/ UX– these tools require immediate time to value; if they do not work quickly, people will never give them a shot‍
5. Insufficient senior push – directives for these tools need to come top-down, and partners need to believe in and reinforce the value of these tools to achieve firm wide adoption

Criteria to Build Internally

We believe the below is the baseline necessary to build a successful internal software tool for IB/PE (note: this not applicable to seed / series A investing, which is more raw sourcing).  

However, many firms fulfill the qualities below and still fail.

Before deciding to build software internally, carefully consider your goals and resources. Building in-house may be appropriate if you need a highly customized, sophisticated tool and have a semi-technical investment professional who can define the project's scope. You must be willing to invest significant time, likely several months, into development and allocate the necessary budget for the project. If you meet all these criteria, internal development may be a sensible choice.

However, many firms build tools without sufficient thought. The most common bad reasons listed below:

• Cost savings – you will almost always end up with an inferior product vs. the cost of buying software
• Data privacy – most software firms offer VPC deployment, which means that everything is deployed within your firm infrastructure (i.e. nothing leaves your VPC instance)
• Experimentation – this will just be a time and budget drain, and you’d learn more talking to experts rather than building in-house.
  • If you want to learn more about what firms are building in-house vs. buying, just email us at info@prosights.co
  • Don’t waste a few hundred thousand dollars and drain your junior team’s time.

Our View

After looking at over 50 internal tools, we’ve now seen too many firms now try to build something in-house without sufficient resources / focus. The outcome is months of time drain, budget, and a tool that your team does not go out and use. We’ve seen this happen dozens of times.

It is ultimately difficult for PE firms and investment banks to build these tools in-house as they are unable to hire and retain top engineering talent. Think about it rationally – if you were a top-notch investment banker, how likely would it be that you would spend your career on Deloitte’s investment banking team? Deloitte is a core audit and tax company – not an investment bank; you would rather go to Goldman Sachs, Morgan Stanley or Evercore.

Similarly, if you are a top-notch engineer, you want to work on the core mission of your company. These engineers typically go to leading tech companies, or AI startups. The crux of IB/PE is advising clients and investing intelligently – not building software. Why would top tier tech talent want to work as an engineer at a core investment firm, where they will not be working on the core mission at a slow-paced investment firm in which they will take a 50%+ discount relative to a high-growth startup with top engineering talent? If you do decide to build in-house and want to hire engineers as competent as those at multi-managers, be willing to match them on compensation (i.e. think $300-$500K for entry level roles). This may be structurally difficult for many investment firms, as your engineers may be compensated better than your junior investment professionals.

In light of these considerations, we encourage investment firms to think deeply about the scope and value of in-house building before embarking on the journey.  If you are a leader of an investment firm, the question you need to ask yourself is whether you are willing to spend several millions of dollars annually with a dedicated on-site technology staff to build this tool, understanding that there is a high likelihood that it will have minimal adoption.

The other option is to purchase a product externally for low 5-figure to low 6-figure ACVs that has been purpose built for your use case. These tools will be tried-and-true, and should work out of the box with varying amounts of set up time. The natural downside here, however, is that these products are typically less customizable and oftentimes are not suited for individual firms’ needs. You will also accrue no competitive advantage adopting these tools – but at the minimum, you will not fall behind. 

Ultimately, the decision to buy or build is not just about technology, but about where you care about maintaining an edge. For VC firms who want an edge in sourcing and have an extremely strong tech-person in-house, it certainly might make sense to build a sourcing tool internally (and this cost will be much lower, likely close to the low 6 figures annually). For private equity, given the role is more process-driven than information oriented, the tools built in-house tend to be more idiosyncratic and bespoke.

What's To Come

In the coming months and years, there will be a new age of LLM-based tools that will allow you to leverage your own internal data to gather more meaningful insights and automate more workflows. It will be at this point when the value of these products will be dependent on the quality of your own internal data / knowledge base. Today, they are focused on basic tasks such as structuring data in tables, producing basic PPT outputs, etc. Going forward, the tools will evolve as the underlying models become capable of high-fidelity reasoning.

That is why, alongside our workflow automation tools, that is why we are partnering with leading investment firms on several bespoke tasks that require complex reasoning. Some examples tasks include:

• Pre-meeting reads – helping partners get a read on the board / stay up to date on latest news, past meeting notes from CRM, and other key data ahead of meetings with external parties (i.e. meeting LPs, other sponsors, etc.)
• Market research – leveraging RAG and crunching through swathes of industry reports into first cuts of what otherwise would be CDD materials that cost $500K+ for McKinsey to put together
• LP reporting – automating the first cut (but ultimately allowing users to edit various inputs / assumptions given the political nature of valuation marks)
• Automating NDA reviews – saving firms low tens of thousands of dollars from having to pay for NDA review for every deal
• OCR at scale – extracting key pieces of data from a large set of unstructured documents into a structured format

If you have any questions, or have any technically challenging problems at your investment firm, please reach out to us directly and we’d be happy to share our learnings or have an initial discussion with your team. 

You can reach us at info@prosights.co 

‍