On leaving Ramp

Last Friday, September 9, was my last day at Ramp. It was by far the best place I've worked in my short career.

The best of times

I joined Ramp last October (as briefly mentioned in my 2021 in Review) as the ~170th employee and the ~30th product engineer.

Ramp, for the uninitiated, is an all-in-one platform for finance teams, and its flagship product is its corporate charge card. For much of its history, Ramp did not support disputing transactions in-app, and instead customers needed to file tickets with support.

A few months prior to my joining, Ramp finally added the ability to dispute transactions in-app and ever since, customers had been disputing an exponentially increasing number of transactions MoM (month over month). This put significant load on Risk Operators to (1) decipher whether the dispute was valid and (2) file the dispute with our card issuing partners via dashboard. To put it into perspective, this was taking 5-10 minutes per dispute, and it was untenable to have this much burden on Risk Ops as the company scaled.

The week I joined, my manager Pablo told me that I would be singlehandedly staffed on solving on this problem and automating the workflow here. The end goal of the project was for (1) Risk Ops to crystallize the SOP (standard operating procedure) around determining valid disputes, and for (2) engineering to build tooling to one-click file disputes. With these combined, Risk Ops would be able to hand off disputes entirely to Support.

I was worried I wouldn't be able to ramp up quickly enough to the codebase and the necessary business context, never mind taking on a project like this solo. Nevertheless, the project was a perfect fit. I was well supported cross-functionally, and I was able to ship the project in a few weeks.

What was next? Pablo (my manager) mentioned a new greenfield project Ramp was considering named "Float". The idea was simple: Ramp had a Bill Pay product where businesses could pay invoices for their vendors directly on Ramp. We wanted to layer on a lending product on top, allowing businesses to get net-30/60/90 day terms. In retrospect, this was a brilliant idea. Ramp is uniquely well-positioned to underwrite businesses, as Ramp has extensive card history + other data (accounting, bank account data, etc) used to understand business health for the card product.

The team was lean from the start – Sam Packard on Product, Jason Li on Design, and three engineers – Max Lahey, Sashank Thukpari, and myself. The initial product spec was written by December 2021, and we began work in earnest after the start of the new year with the ambitious goal of shipping a MVP in March.

After two months of long-running Zoom meetings, shipping code as quickly as we could open pull requests, and a late name change to Flex, we shipped the MVP in mid-March. It wasn't obvious that Flex would be a success from the start. We outbounded to about X businesses (number hidden intentionally), and although we we were able to sign up a sizeable percentage, businesses weren't actually using the product and initiating loans.

We suspected that there could be a few reasons for this. (1) Flex had no PMF, (2) variability in when businesses sought out loans, (3) the interest rates we offered were not competitive enough. We sprinkled a few growth hacks for Flex throughout the product, such as interstitials deep-linking to the product or banners within the Bill Pay product.

After a few more weeks, our patience paid off - we had our first loan initiated on Flex. Over the next few weeks, it became clear that the product was incredibly sticky for businesses using the product, and that it was worth investing in this greenfield project further.

We ambitiously set a target date of August 15th to launch the product for GA (general audience), and began working on the necessary feature set to release to everyone. As Sashank noted in his twitter thread, this was a major cross-functional effort. It was my first time witnessing the level of coordination required for a large product launch, largely driven by Subham Agarwal and Serena Mackool. Ramp invested more engineering support for the GA launch, with Arnab Dey and Jack Chau (intern) contributing.

Despite some last-minute hiccups, we launched the product on time and bug-free (thankfully). Here's the blog post on the product launch for those interested.

I'm truly honored to have worked with such a brilliant group of people to make the launch happen. Although I'm biased, I wholeheartedly believe that the core team that shipped Flex is the most talented team at Ramp.

Reflections on Ramp

It's certainly bittersweet to say that I'm leaving Ramp. There's definitely a world where I stay there for the rest of my career. Although no company is perfect, Ramp is far closer than any company I've seen.

I've been asked what made Ramp such a special place to work. I think like everything else, it always comes back to the people. Ramp is incredibly talent dense, and especially so in product/engineering/design. Assembling talent of this caliber is already rare, but it's even rarer to have a culture where everyone is highly motivated by what they work on.

During my time, I was particularly inspired by two people – Max (TL on Flex) and Sam (PM on Flex). Max is incredibly thoughtful and humble for someone who is correct nearly all the time, and I was awed by his ability to take complicated problems and eloquently communicate concrete paths forward. Sam is detail-oriented and can absorb information at an unbelievable rate. We used to joke at Ramp that Sam could be replaced by sentient AI, and we wouldn't be able to tell the difference. I'm certain they will continue to bring the best out of everyone they work with.

Karim/Eric/the rest of the leadership team have done an unbelievable job bringing Ramp this far, and I'm excited to finally be on the other side as a Ramp customer.