Categories

Startups
Startups

How we Built a Startup Fundraising Network that Aims to Remove Bias

The DFN has been in the making for almost a year now and it consisted of three separate MVPs that we iterated on to get to the offering that we just launched. Here’s the story of how we did it.

On June 25th, we formally launched the DocSend Fundraising Network (DFN). This is something we have been very passionate about here at DocSend. Our goal with the DFN is to surface meaningful new deal flows for VCs while helping founders who have quality ideas, but might be under-represented or without easy means to get in front of the right investors through current channels.

As you will see, we are trying to use a data-driven approach to identify quality startups that deserve VC attention and remove some of the common biases in the fundraising process. But, to be honest, that’s not how we started. The DFN has been in the making for almost a year now and it consisted of three separate MVPs that we iterated on to get to the offering that we just launched. Here’s the story of how we did it.

It All Started with Our Research

Over the past year at DocSend, we’ve spent a lot of time and effort trying to understand what makes a successful startup pitch deck. This was at the heart of our content strategy that centered on creating data-driven content that was unique and immediately relevant to some of our main target audiences of startup founders and VCs.

In Q4 of 2019, we surveyed hundreds of founders and VCs about what factors helped them get intros and what factors led to securing term sheets. We developed a survey of over 130 questions with multiple flows and tried to get as much information as possible from our community of users who were kind enough to share their decks and fundraising experiences with us. At the same time, we also aggregated usage data from over 100K pitch decks and looked at tens of millions of data points from the DocSend platform. We then loaded all that data into our data warehouse (AWS Redshift) and went to our BI tool (Periscope, now Sisense Cloud) to analyze it. Investing in a proper data infrastructure (vs doing our analysis in spreadsheets as we were initially inclined to do) was a key early decision because this setup gave us agility and speed to explore our data in different ways, which became very important later on.

Through our data exploration journey, we were able to test commonly-held assumptions around fundraising. Is it good if a VC spends too much time on your deck? Do I need a technical cofounder, or can I raise on my own? What are the most important factors that VCs weigh before agreeing to meet with a founder? Is there a glass ceiling in the amount of capital raised for women founders? In every case, we tried to pull actual data and challenge what everyone claimed at face value.

In Q1 of 2020, Andrea Silvers, our Head of Content & Comms, and 104 West, our specialized PR agency, published the first set of our research: The Pre-seed Round Defined: How to Succeed as an Early-Stage Startup (downloadable here). This research pulled together all our learnings about what makes a good Pre-seed pitch deck. The report very quickly got thousands of downloads and sparked more than a few social media conversations. One of our key findings was the following diagram that was particularly striking.

From Research to Methodology

Looking at this diagram, we felt we had in front of us a practical blueprint of what can be a good deck and what can not. And this blueprint was totally objective. It didn’t take into account the race, gender, or background of the founding team. It didn’t require a VP or an Associate in a VC firm to process the deck with their own set of criteria. So, fueled by this diagram, we put our research findings to work. We created a rubric to score Pre-Seed pitch decks. The rubric had nine data points with different weights and each data point could receive a score from 1 to 5. The overall score we assign to a pitch deck is a percentage between 1 and 100. We also built a detailed feedback template that would dynamically follow the scoring from the rubric and provide specific, actionable suggestions of improvement for the deck.

We then set out to test this scoring methodology and the dynamic feedback that followed it. We collaborated with a number of founders that allowed us to score their decks and see if the feedback made sense. We sat down with some of our partners (like Atrium) that also worked with early-stage startups and got their feedback. We referenced our results with VCs (like Kyle Lui from DCM or Aileen Lee from Cowboy Ventures) in our network and got very constructive input that challenged us to think harder. We even used our new methodology to score some well-known pitch decks (examples here). Through all these interactions, we were able to iterate on our rubric (which we expanded to 11 data points) and gain confidence that this was something that: 1) clearly seemed to work and 2) could be genuinely useful.

In late Q1, we quietly launched our first MVP: the DocSend Pitch Deck Analyzer. Anyone on the DocSend platform, whether paying customer or not, could send us a link to their Pre-Seed deck and we would analyze it and provide actionable feedback. The Pitch Deck Analyzer “launch” consisted of a web page and an email address. We didn’t much know what to expect. And then we started receiving decks, lots of them.

To keep up, we had to build a process to receive these decks properly, ensure privacy and confidentiality with the founders that shared their decks with us, and aim for a 48-hour response turnaround time so that founders were not left hanging. And we committed that every deck score would be verified by a human. The team (this was ran by the marketing team at DocSend) was stretched. But the feedback we were receiving from founders was phenomenal.

Such a great initiative, love this! Thank you guys for doing it.

Thank you for the analysis. It was surprisingly detailed with good pointers made. Appreciate it.

That’s the most constructive feedback I’ve ever seen!!! Thank you for that! I’ll work on all mentioned points, once finished, is it ok to reapply?

This was incredibly helpful! I love that you offer this service to founders.

And what was interesting was that the feedback came from so many different and diverse founders and businesses. Looking at this feedback was the moment when we realized that this methodology could help alleviate the three most common biases used in early fundraising: Who are you? Who do you know? What’s your network? Maybe we could do something different here. We were getting excited!

Getting Back to More Data

And then, the COVID-19 pandemic broke out. We quickly decided to shift focus onto something more timely with our data. We started looking at real-time data from the DocSend platform and launched a set of Pitch Deck Interest metrics that gauged how supply and demand of pitch decks (and consequently, ideas and investments) were being affected by the pandemic and ensuing economic slowdown. The fact that we had our data in a data warehouse and could easily explore it using SQL allowed us to move fast.

We made our metrics public here, we updated them weekly and we tracked that despite an initial slump in March (primarily due to the level of uncertainty), both startups and VCs adapted quickly to the new reality. The startup fundraising market was quick in getting back on its feet, albeit with one major change: face-to-face meetings were not possible anymore and the proverbial drive to Sand Hill road lost a lot of its allure. While the Pitch Deck Interest metrics received very good attention from tech media publications (like this article in TechCrunch), for us they were our second MVP as they proved to be accurate leading indicators of how the fundraising market moved and this learning built further confidence in the quality of our data and the signals we could generate. We also realized that if startups had to run most of their fundraising process remotely, the Pitch Deck Analyzer could be even more useful.

The Other Side of the Table

As we were seeing the fundraising market recover, we found out that some VCs were faster than others to open up to virtual meetings and try to build trust with founders remotely. This led to our third MVP, the DocSend Active VC List that we crowdsourced with our community and made available here.

When we started the list we were struggling to find enough VC entries to make the vertical scroll bar appear at our Airtable, but very quickly we got to over 300 VCs that were very open with us as to how they were adopting their processes, when they were becoming available for meetings with new companies, and what was required for them to restart signing term sheets. The Active VC List also gained a fair amount of attention (like this article in Business Insider). Through this effort of connecting with VCs and seeing how they were approaching fundraising in a different world, we realized that we were actually sitting squarely in the middle of the table: on one side we had a great methodology to identify quality companies through their pitch decks and on the other side we had relationships with VCs alongside their investment preferences and signals on how quickly they were moving. Everything clicked into place.

Putting Everything Together

With a final push by Russ, our CEO who had been super supportive and enthusiastic throughout this effort, we set up (quietly, once more) the DocSend Fundraising Network pilot. We quickly scaled to over 300 pitch deck submissions and over 30 VCs joining this pilot program. We had to go out and hire a dedicated resource to help us run the network. While keeping a high bar on quality, we started to see the first matches and meetings taking place. And we still provided (and continue to provide) detailed feedback to every single deck that was submitted to join the DFN.

We are excited to formally launch the DocSend Fundraising Network today. We are focused on Pre-Seed and Seed rounds, but we plan to expand to Series A rounds soon. We are equally excited about next steps. We believe fundraising is something that can be approached differently. We hope that this effort will provide some meaningful value to under-represented entrepreneurs and create a more inclusive, global network of founders and investors. Although it was not our initial goal, we learned a lot about fundraising bias from our journey so far. We will continue to learn.

PS It literally takes a village to launch something like the DocSend Fundraising Network. It’s hard to name everyone who has contributed, but special shoutouts to RussDave & Tony, DocSend’s founders, for their guidance and support, Andrea for her relentless passion and commitment to making the DFN a reality, Bryan and Becky for setting up the data infrastructure, Alessandra for building up relationships with VCs, Jenny and Nick for getting the word out about the DFN, Chris for filling in numerous holes with our messaging, Sam for automating a lot of our back-office processes, Meera for coming in fresh after her MBA to help us scale this effort, all the founders and VCs who gave us feedback along the way, and many more.