<span id=Pinpoint’s Year in Review — Looking back at 2019, Forward at 2020" />

2019 was a big year at Pinpoint. Let’s start with some of the obvious successes: We started the year with a great Series A funding round that was announced on the same day that we launched our first product. 

As a technical CEO working with a very technical team, and supporting technical customers, I’m deeply invested in the strategic decisions about what features to add to the product, architecture design improvements, and how to elevate our customers’ experience with Pinpoint. 

Here are some of my takeaways from 2019, both from a business and technical perspective and as a user of Pinpoint and the impact it made on how we work.

As a side note, because we value transparency, the screenshots featured here are our actual Pinpoint dashboards, showing real performance. 

Invest In Things That Matter

After launching the product early last year, we started using Pinpoint ourselves. We started Pinpoint because of our own frustrations measuring engineering performance and always intended to use the product from day one. One of the first things we noticed was that our average Cycle Time for new feature development was 102 days — way too long. We needed to find a way to move faster. Pinpoint allowed us to further investigate what areas we needed to focus on in our process. 

Pinpoint Cycle TimeFigure 1: High Cycle Time was indicative of a significant issue with our process that needed to be investigated.

The bottleneck we found was in our backend architecture of how we managed our data processing pipeline and how teams had to work in serial mode for any major changes to the pipeline or data model. Any change to the data model took hours to process existing data and lots of coordination between team members, which drained our development team’s productivity and made relatively small changes difficult to execute.

Duration by type

Figure 3: Long duration times impacted our ability to iterate quickly.

This discovery prompted us to completely change our data science pipeline and application architecture. Specifically, we moved to a real-time data science pipeline and a more document-oriented data design. Our team redesigned the cloud agent to be more flexible in the data schema to allow changes to evolve more quickly and migrated to Kafka. We also built out a more formal data science pipeline and stopped running ad-hoc calculations for complex values by moving them into the pipeline.

The good news is that these changes made a significant difference in the way our team developed and released software. We’re now able to develop new features much more quickly than before, delivering now in days, not weeks. Processing speeds are much faster and our team isn’t wasting as much time waiting for changes during the development cycle. The downside to this exercise is that this initiative took six months longer than anticipated and required reworking the entire data pipeline and data storage layer from scratch. 

Pinpoint Cycle Time improvement

Figure 3: Much improved results for our company over the past 90 days.

The end result is a product that is much easier to scale and that allows us to develop at the speed our teams and customers expect. Most importantly, it allows us to respond quickly to customer feedback — a goal that I believe is shared by any SaaS company. I’m pleased with where the product is now and think the massive investment in redesign was worth it. 

Building a Team To Last

Now that our team has more than a year working together, we’re starting to develop institutional knowledge and a positive workplace culture. We started 2019 with just a handful of developers and ended the year with around 20 developers. There’s also flourishing cross-team development now that we have more visibility into how, and how well, we work in various areas of building the product, which I think will contribute to the health of the company long-term.

Goals for 2020

My larger vision is for Pinpoint to help the people who build software, build software better. It should be a platform that developers, managers and executives can use on a daily basis to understand what work is being done across teams, projects and systems to improve efficiency and uncover new bottlenecks in their software delivery. Our focus has always been on providing information that’s actionable and backed by trends in the data — we don’t want to just visualize data for the sake of visualizing data.

I also want to double down on approaching our own use of Pinpoint as ‘customer zero.’ We’ve been able to develop a very good intuition for what is good, and what needs improvement in the product based on our own team’s everyday use. Going forward, it’s going to be important for us to continue iterating quickly and building new features based not just on our customer’s needs, but on the value and pain points that we can also identify as engineers. We will continue to “eat our own dog food” with each new feature, while feature flagging quickly to end users for direct feedback.

Project Dashboard - Current Sprint - App Engineering

I also plan on opening up the platform by offering an SDK for third-party integrations and an API to interact with data in Pinpoint. By doing that, we will greatly expand the ecosystem around Pinpoint and make it easier for customers to tailor the product to their exact needs and to add additional value on top of Pinpoint. 

2020 will be the year we continue to prioritize building the features that engineers find useful. Together with partners and users, we’ll change the way that we optimize the process of building software. 

Get the data science behind high-performance teams.