Investment Series: PagerDuty, Disqus, and MemSQL

Anish Dhar
Jan 26, 2017 · 5 min read

Continuing on my #dealmemo investment series, I’ve been thinking about three companies I would invest in today if I could. These are more growth stage and all have raised multiple rounds over the last couple years, but I strongly believe there is still monumental upside and impact left to be made in each market they reside in.


On any given day at Uber, if you walk around long enough you’ll eventually hear this loud robot yell “PAGERDUTY ALERT” about a million times on someone’s phone, eventually convincing you that PagerDuty is probably some horrible practical joke used to mess with engineers. While the call itself can be annoying, it definitely serves its purpose, as PagerDuty has dramatically improved response times to major outages at Uber by sometimes 10–15 minutes, which amounts to millions of dollars over time. In fact, if you ask any engineer at some of the world’s fastest growing startups like Uber, Airbnb, and Okta what the most important tool they have at their disposal to ensure reliability, you’ll most definitely always receive PagerDuty as the response.

PagerDuty seems like such an obvious solution to a problem that tech companies have been dealing with forever, but has only gotten worse as companies have adopted the Facebook mantra of rapid continuous deployment. When you have hundreds of teams who are all making daily changes to a codebase, it becomes almost impossible to build custom protocols that will alert stakeholders when their systems break. PagerDuty provides a clean centralized system that has strong integrations with products like Slack and Jira, making it an invaluable part of any growing tech stack.

The beauty of PagerDuty and why I believe its here to stay is because of the increased reliance a tech company has on the product as it scales. As PagerDuty becomes used to power reliance on more important services within a tech company, engineers develop stronger trust in it and give it more responsibility as they continue to build out the company, making it one of the most powerful tools in terms of sheer control. I’m convinced that PagerDuty will become the de-facto tool every growing tech company will use to monitor their services, and one day will even be able to automatically fix common problems it alerts engineers with. When that day comes, you can be sure PagerDuty will become one of the most valuable enterprise companies in the world.


Growing up, I would often spend a lot of time browsing YouTube or other forums and reading through the comment’s section, sometimes even responding to a comment I thought was funny or insightful. Yet as years passed and these services became increasingly popular, the comment’s section often involved into spam or even sometimes just plain abuse, with little moderation and no real sense of community. That’s why Disqus has achieved something incredible with its discussion product: a sense of real community throughout the internet.

The best companies start off with a core group of users who absolutely love the product, and Disqus is no different with sets of power users who have left thousands of comments on hundreds of different websites over the years. It has now evolved to millions of users leaving comments across the web, yet Disqus has done a remarkable job scaling and keeping engagement high among discussion groups. Disqus content is often thought provoking, and users feel motivated and engaged within the product creating a strong sense of community within each comment section. I’ve seen the same users interact with each other across the internet on different websites, which is the real beauty of Disqus and why I believe it will grow to a multibillion dollar company one day.

Not only do users feel engaged, it’s clear that publishers see the value of Disqus as a tool to grow their own subscribers and keep them coming back to read their content. In fact I believe that in the future page views will be almost less important to a publisher versus the engagement they see on their story through platforms like Disqus, where users could engage for hours on some content that has ads powering the platform. It’s really interesting to see how users who use Disqus leave their imprint on the web, and it’s almost like annotating the internet. I strongly believe Disqus will become an incredible social platform in its own right, with publishers signing on instantly because of the unique value it brings to any website.


As more and more services continue to grow into real time platforms that ingest tons of data every second, performance becomes incredibly crucial and expensive at scale. MemSQL is literally optimized for this specific use case, and solves many of the problems traditional SQL / NoSQL databases have to deal with in real time data ingestion.

By combining a traditional MySQL layer with the best features from NoSQL, you optimize for all the inefficiencies that plague some of the more standard databases like Postgres or MongoDB. Working at Uber, I’ve seen first hand how slow database analysis can become when dealing with millions of read / write functions occurring every second on a traditional SQL database. Yet the speed and efficiency of MemSQL is not why I believe it’s going to become an incredibly significant company in the world of data analysis, but more so its strong integrations with products like Kafka and Spark.

Spark is an open source data processing engine that powers the real time streaming analytics at some of the world’s most important companies like Netflix and Pinterest. MemSQL is one of the few database companies that has heavily optimized its platform to integrate with Spark, and has built business intelligence tools to empower users to easily use Spark in conjunction with the database. It has done similar projects with Kafka, another open source tool used by thousands of companies to log messages and events to servers. You can see how companies are starting to already take advantage of MemSQL’s extremely efficient integrations going to any big data conference or talk, where dozens of companies like Pinterest have publicly spoke about how fast their streaming analytics became after switching to MemSQL.

I strongly believe that MemSQL will provide the database layer of the tech stack as companies have to deal with more real time analytics or generally any big data processing. As tools like Spark / Kafka become increasingly popular in tech companies, so will MemSQL as it continues to prove how dominant and useful it can be in performance.

Anish Dhar

Written by

Engineering / Product @ Uber, previously summer analyst @ NEA.