Rainbows, Unicorns, and Spaghetti
What I Learned From One Year Researching Startups at a Silicon Valley VC Firm
Let’s play a game called “Guess My Grade.”
Your job is to guess what grade I got on the assignment:
The teacher left as many words of feedback on my essay as there were words in the essay.
What grade do you think I got?
Yup, not looking good.
Well, on my first ever draft of a Contrary Research memo, I received 1399 words of feedback across 35 comments on a memo just 1400 words long. I wasn’t exactly off to a great start.
Everyone’s experienced a moment like that before. They were excited for a new opportunity but quickly realized they were in over their head. That’s how I felt when I started writing for Contrary Research in May 2022, right around when I graduated from college. I had been exploring the world of startups over the past few months and somehow managed to get accepted to the Research Fellowship. I continued writing for almost a year until I started a new job as an engineer at Maybern (a seed-stage startup). After 8 memos, one 7K word deep dive, 36K total words, dozens of edits, and a bunch of new connections, I wanted to pause and reflect on the experience. That’s what this post is about.
My feelings towards the Research Fellowship followed a hype cycle-like trajectory. When I was first accepted, I was super excited about the opportunity to write and publish while learning about startups (”Rainbows and Unicorns”). I quickly realized I had a lot to learn (”Uh-Oh, SpaghettiO”). I put in the work over many months to improve (”Lessons Learned”). And now, I can take a look back at my experience (”Looking Back”).
Rainbows and Unicorns
Contrary Research is a website run by the venture capital firm Contrary that aims to the best starting place to understand any private tech company. Since launching in September 2022, they have published in-depth pieces on well over 150 companies.
In April 2022, Contrary Research was recruiting its first cohort of writers, which they called Research Fellows. I had connected with members of the Contrary team, including Will Robbins and Kyle Harrison, while searching for software engineering opportunities at startups. A few days after chatting with Kyle, I happened to see his LinkedIn post about the program. He hadn’t even announced he had joined Contrary, and Contrary Research was pre-launch and in stealth as they were starting the program.
I figured I could learn a lot about startups while developing my writing skills and gaining online exposure. After an initial application, an interview with Francis Odum, and a few follow-up emails, I finally received an offer to work with the team. A slight complication was that I was going to be traveling internationally for much of the summer. But given that the gig was part-time with flexible hours and provided a unique learning opportunity, I happily accepted.
Uh-Oh, SpaghettiO
When my very first draft came back smothered in 1399 words worth of comments, I was disheartened. I had spent a solid 20 hours reading up on Persona and writing what I thought was a decent memo. Looking back on it, the draft was full of generalities and a few startup terms I was vaguely familiar with. Below are three lowlights from my first draft:
The identity verification market is large and growing.
Low Switching Costs: Customers could relatively easily switch identity verification providers and be OK
[Persona’s] $1.5B valuation seems reasonable given the large and growing identity verification market and Persona’s strong execution thus far.
Large and growing market? → That applies to most VC-backed startups.
Low switching costs? → Ok, but explain why.
Strong execution? → How so?
The Contrary Research team maintained high standards from the start. Although I certainly found the revisions to be painful at the time, there is no denying that the end result, with a heavy dose of editing, was much improved. For example, here’s the current section on low switching costs:
Low Switching Costs Customers can switch identity verification providers without tremendous difficulty. The co-founder of one existing Persona customer, MyRent, described how such a switch would face three surmountable barriers: (1) Continuing support of Persona until existing users upgrade to the latest version of the app, (2) Exporting data from Persona and mapping it into the format used by the new vendor, and (3) Integrating the new vendor within the app.
MyRent’s co-founder estimates the transition would require ~3 months of engineering time, with one month for a backend engineer and one month each for an Android and iOS engineer. In addition, waiting for existing users to upgrade would take approximately 6–9 months, during which time the company would have to pay both Persona and the new vendor.
Persona’s vision of becoming more critical in the identity stack, and expanding to manage user data, would increase the switching costs for their customers.
Lessons Learned
With enough time and hard work, I was able to finish a few reports. I slowly distilled three key lessons that are most applicable to research memos but are also useful to have in your writing arsenal for other styles.
1. Back It Up
For any claim you want to make in your writing, back it up — and if you can’t back it up, then delete it. The evidence can be anecdotal or statistical, but there’s got to be something. This lets anyone reading the piece understand how conclusions were drawn. There were often times when I was writing where I got the “vibe” that something was true, but looking for evidence forced me to clarify and challenge my opinions. To see what backing up claims looks like in practice, try counting the number of hyperlinks in my Stripe deep dive (I counted 336!). Also, see the comparison below of a “Key Opportunity” I wrote about in my first draft for Cohere, which bordered on pure speculation, with one I wrote for Stripe, which was backed by data and details about Stripe’s competitive positioning:
Cohere
Another monetization opportunity for Cohere is analyzing the API requests it receives to identify trends. For example, it might notice that consumers using its customers’ chatbots are asking more about certain topics. These insights could then be sold to people who find the information valuable (e.g., marketers).
Stripe
Only 3% of the world’s companies transact outside their country’s borders as of November 2022. Zooming in on the US, Stripe’s home market, just 1% of SMBs export abroad as of 2017. This could change as technology eases the friction required for international operations.
Stripe contributes to international commerce by supporting payments from 197 countries and a few dozen payment methods. Although this coverage trails some large competitors like Adyen, it leads smaller domestic players such as Razorpay and Xendit. They lack support for relevant payment methods outside their home markets, including Apple Pay, Pix (Brazil), and iDEAL (Netherlands). In comparison to these more limited offerings focused on local markets, Stripe is a better positioned for companies that want to sell globally. This is especially true for SMBs, which competitors like Adyen don’t cater to.
2. Tell A Story
While the reports published by Contrary Research aren’t necessarily meant to take a strong stance on a company, they aren’t a simple regurgitation of facts either. The facts need to be linked together into a broader story. It’s the difference between having a pile of scattered, individual Lego blocks versus combining them into a useful, cohesive structure. As an example, here’s a look at the second paragraph of the Customer section of the report I wrote on Stripe:
The steady expansion of Stripe’s customer base was reflected in the sequence of its product offering launches. Stripe launched Connect in 2012 to serve platforms and marketplaces like Shopify and Lyft that need multi-party payments. In 2018, it launched Billing to handle the needs of subscription-based SaaS companies such as Figma, Slack, and Zapier. Stripe supports invoices and introduced Financial Connections in 2022 to ease payments for bank accounts, which are important for B2B companies’ payment needs. B2B customers include Atlassian and Maersk. These adaptations have allowed Stripe to meet the needs of an increasingly wide range of businesses.
This paragraph shows how Stripe gradually expanded its customer base to include different categories of businesses rather than just stating that Stripe serves various sizes and types of companies.
3. Keep On Digging
The research process for these companies can take a lot of time and requires digging beyond the surface. Contrary Research provided access to valuable paid resources like Tegus, Sacra, and Crunchbase, although I still heavily relied on Google. I would often try multiple variations of a search before finding what I was looking for. Sometimes, I’d accidentally come across new evidence, resulting in updating or adding to a section. That could be frustrating at times, but the patience always paid off in the end.
One area where I found digging into the details to be illuminating was when understanding the differences between Stripe and its most direct competitor, Adyen. I initially just wrote that large global enterprises seemed to prefer Adyen over Stripe and cited a source for that. But eventually, I came across resources that provided a better explanation of the dynamics, resulting in a better understanding not just for myself but for my readers. Here’s a snippet from the Competition section where I summarize a few key differences between Adyen and Stripe’s products.
Adyen’s offerings have been tailored to global enterprises’ needs. Stripe’s point-of-sale solution, Terminal, lags behind Adyen’s, although Stripe’s acquisition of BBPOS may help address this. This helps Adyen win omnichannel deals with companies like Subway and McDonald’s, even as it attracts digital enterprises like Uber and Spotify. In addition, Adyen provides support for more countries and payment methods than Stripe. Adyen supports over 300 payment methods in nearly 100 countries compared to fewer than 50 payment methods in 47 countries for Stripe.
Looking Back
As I gained more experience (and confidence) writing the memos, I tackled increasingly high-profile startups, including Flexport and Brex, and finished up with the giant that is Stripe. I went from receiving feedback that made my heart sink to my Stripe deep dive being featured on the Contrary Research homepage.
Another important aspect of being a Research Fellow is getting to know the rest of the team. I’ve been lucky enough to be able to lean on multiple people I’ve met through Contrary for job advice. A few of my favorite memos from other Research Fellows include Replit (by Nick Bello), The Boring Company (Mike Kelly), and Gong (by Allison Kiang). I’ve read a bunch of memos myself and have learned a great deal from them. A big thank you to everyone who makes Contrary Research possible! If you’re interested in being a Research Fellow, you can apply here.
The Research Fellowship was a unique and overwhelmingly positive experience for me. It was helpful when I was looking for an engineering job, as it showed startups that I was genuinely interested in the world of startups and had business sense to go along with my technical skills. I developed both my knowledge of startups and writing skills while building connections with great people — all of which I will take with me throughout my career. And coolest of all, I now have a public portfolio of writing that I’m proud of and that people actually read. For now, I’m focused on my engineering job at Maybern (which is terrific!), but I hope to come back to writing at some point in the future.
This is an awesome read Rohan, done some similar work for a venture firm before, it's def exiciting :D
I'll probably try getting into the fellowship sometime haha!