Deal flow to investment decision. Can we automate early stage investments?

The top investor reviews about 12 investment opportunities daily. From a generation stand point, these investors receive a high level of inbound deals,  on top of the ones they generate for themselves. The rest of the pack have to step up their outreach efforts to be able to generate enough deal flow.

How do VCs manage 12 opportunities per day, 50 weeks per year, 5 days per week? Screening 3,000 deals is a major effort which requires fine tuned processes and good access to data.

We have seen that venture capital companies have a largely manual deal flow process based on scout and referral programs, reaching out, exchanging hundreds of emails, analyzing decks and trying to centralize everything by hand before making a decision. 

One angel investor was detailing on a quora thread the level angels and micro VCs are probably at. “I know that I personally look at 200 deals a year, narrow the field to 20 and select 2. That is a rate of 1%.” Even if more were available, would he be able to pre-screen and evaluate those deals by himself?

Startup investors are running mainly a referral game. For some, it all depends on the strength of the network they reach out to. Jeffrey Glass, a partner at Bain Capital Ventures, explains that “In one year I see over a thousand business plans and meet on average with hundreds of companies, but ultimately only invest in one to two.” For others, it is a matter of how well they use scout programs and other available resources. 

While venture capital is an industry that spearheads innovation, the whole investment process is still incredibly slow, labor-intensive, inefficient, expensive, and often even biased (Sheehan and Sheehan, 2017). Investors still choose to pump funds into the ideas of friends or friends of friends. They make decisions based on intuition rather than data. But intuition or gut feeling is simply our brain’s analysis over historical data points and on the fly risk-reward calculations. Maybe the industry could do with some additional, data-driven digital “gut-feeling” processes. Deal flow automation & due diligence digitization could be first steps toward a better digital investment platform. 

Can automation work in startup investments?

If we lay out the basic investment process we would be looking at some general steps.

Sourcing, screening, and review

The main focus and effort goes into developing a strong referral network (other investors, portfolio companies, scouts, etc), scanning startup websites and product reviews, browsing recent news and interviews, and reviewing investor decks where available. The process is mostly manual and time-consuming, and there is little data used in this step. Some companies are surely spearheading efforts in this area. They understand that efficient deal flow growth needs digitization, automation, and data. Sanjiv Soni describes it in this article.

First meeting/touchpoint 

Once the preliminary analysis is done and green flags pop up, the process moves into meeting requests, scheduling, and the usual courtship dance. This is a step that validates or invalidates the initial assumptions. The energy with which the founder presents himself, the team, and the business will lead to a positive or negative result. A positive result is what allows for the next step to take place. 

Initial Due Diligence 

This part of the process is data intensive and, hopefully, data-driven. Business model KPIs, growth metrics and financials are usually Excels and spreadsheets, and they are filled with unstructured information. Which makes it hard to extract relevant data.

However, several areas (outlining the marketing strategy, detailing partnerships and supplier relationships, team insights) do not depend on data. Thus, creating and maintaining the deal room information from investor to investor can be difficult for founders.

On the investor side, this information needs to be cross-checked with internal data about markets, similar investments, past investors’ performance, and others. This stage is all about collecting, sharing, normalizing, and analyzing different sets of data. (bell rings…and rings)

Term sheets & Contracts

This process consists of back and forth communication and negotiation of terms. Sure, it can be optimized. But I think that making this step more efficient is only a matter of ensuring that Screening, Pitching, and Due Diligence become more data-driven. If everybody trusts numbers and processes more, fewer risks need mitigation, and the actual negotiation can be reduced.    

Automation bells silenced by habit and process

I am a big fan of automation, and deep down I believe it will make the world a better place. Especially the investment world. And if not all,  at least some parts of the deal flow journey can be automated. 

Automating sourcing and screening can have a huge impact. I know there is a case for deal flow automation at every stage of this basic flow. 

The main issues I see revolve around these points:

  • There is little consensus on data standardization for startup investment processes.
  • Everyone wants more insights. However, few are willing to share and keep to themselves, building silos of data that are never complete, and some never truly relevant.
  • Founders need to adjust and beautify data as investors want to see the risk-reduced, dreamy Disney story for investment rather than a risk-engulfed, but potentially positive pattern of stable growth. This leads to…
  • Founders do not trust investors with their data. They prefer (and are also advised) to share only the minimum necessary for the investor to sign the check. Investors do not share data with founders (due to lack of time, trust, or proprietary considerations). Investors do not trust other investors with their data either. Nonetheless, they rely on co-investing and a general historical performance validation of the co-investor as decision making criteria.

One of the most important things for a VC today is getting to startups before everyone else does and before the market becomes saturated. This is why some have chosen to defy the status quo and use a data-driven approach. Some VCs use existing data platforms for this approach. Others have developed proprietary software or databases to gain a competitive advantage.

“It’s to get companies on our radar early because there are only so many companies we can see, so having more companies helps widen our aperture,” said Lightspeed partner Jeremy Liew

If you were to remove the blockade and make technology work 

Some investors are more transparent than others about the software they use or create and how this software helps them. Others avoid specifics on the subject or avoid the subject altogether. It could be because they fear losing their advantage, or there is no direction for deal flow automation at all. In reality, there are several ways to build at least parts of an automated system that works for investments. Data sources and specific metrics that best fit your fund’s philosophy can be tracked to some extent with what we have today. There is no out of the box way to do it, nor a blueprint, but it can be done.  

Where can you use automation?

Trend discovery. Identify applications and technologies that are likely to become viral hits well before they reach critical mass, as well as trends within your portfolio or network.

Founder discovery. Some signals identify founders early on, even before they become founders. Like everything else, you can use pattern matching here with some success.

Founder analysis. Track potential founders based on predefined characteristics, as well as their digital footprint. Such predefined characteristics are companies they worked for, schools they attended, people they know.

Early stage startup discovery. Identify startups early on, as early as MVP or proof-of-concept phase.

Sourcing automation. You can use this to connect with a larger number of startups early on, stay in touch, and nurture a relationship. Sometimes, you can do this even before the startups have begun looking for an investment.

Track growth. Track startups and perform due diligence even when there is no active investment process.

Data management. Collect and maintain basic data across different startups, markets and stages. You can use this for benchmarking as well as identifying investment opportunities and easing due diligence.

The process is as basic as making toast. You discover and collect information about potential founders and startups. You enrich that information, track it over time, and store it for analysis.

The challenge is, as pointed out earlier, that everyone holds on to their data.

Using available technology for deal flow automation

You can build some automations for founder and startup discovery by monitoring main startup marketplaces and using scraping tools to build an initial and up to date database. Some of these platforms allow for API integrations which makes it even easier to create the initial data set. 

You can automate founder tracking and startup tracking a bit more with google search rss feeds, building at least a timeline of activity for the person or company.  Another potential workaround is LinkedIn scraping of individual profiles. Though, it must be said that LinkedIn has been trying to reduce this potential. Either way, if you can automate or optimize how you connect to potential targets on Linkedin, you can then use your data export as a regular update for founder status.

For further integrations, Zapier, Integromat, IFTTT as well as several other workflow automation tools can move data around from scraping tools to CRMs and databases. Once a database is taking shape, you have several options for tools that can improve company research results and info. Such tools are LeadGenius, DiscoverOrg, Openprise, and FullContact. Startup investment is a B2B business, and investors could take a few points out of the playbooks of the businesses they fund. Ed Fry from hull.io has written a great article that provides the basic insights into how B2B companies improve their sales data.

Integration now, a shared investment platform in the long run

Designing a deal flow automation model that relies more on market integration could allow investors to evaluate significantly more businesses, from more geographies and markets. This would improve the investment process from discovery and sourcing to decision making. 

The more investment opportunities investors can be exposed to and the more structured data can be collected and analyzed during the process, the more can the decision process evolve. Automating deal flow translates into easing processes, developing more leads, increasing volume and improving quality. Alex Graham wrote on the importance of why “VCs need to get to the front of the next information advantage”.

As the industry evolves, so should investors. Right now, we can see a trend of individual and proprietary software being developed. Bartosz Trocha points this out  in his article about data-driven VC efforts. At the same time, generic startup marketplaces become more popular and the norm for scouting new startups digitally (AngelList, F6S, Gust, etc). 

This is very similar to the trends of more mature markets where OnPremise developed into Iaas, Paas and then Saas models. We are still looking for a true Saas investment model where applications and data can be shared and used at their full potential. 

Think of the Facebook advertising algorithm. Advertisers pay for data collection, data which the algorithm uses to understand and decide where the investment of said advertiser can provide ROI. Thus the algorithm delivers a pretty accurate ad to the advertiser’s matching persona. A similar mechanism can be developed at the tech startup level. This would benefit founders and investors equally, and provide room for meaningful conversations to happen faster and easier. Of course, such an algorithm can be skewed and biased, but this is a challenge to be solved in the long run.