Key Moments

Startup Technology - Technical Founder Advice

Y CombinatorY Combinator
Science & Technology4 min read60 min video
Oct 10, 2018|120,625 views|2,375|76
Save to Pod
TL;DR

Technical founders share insights on building V1, balancing speed with best practices, and team structure.

Key Insights

1

Prioritize speed and iteration for V1; robust processes can be developed later.

2

User feedback is crucial; build prototypes to test ideas early and often.

3

Technical debt is inevitable; manage it by prioritizing core functionality and refactoring strategically.

4

Embrace flexibility in engineering methodologies, adapting them to the company's stage and needs.

5

Hiring full-time employees for core competencies is generally preferred over contractors in early stages.

6

Remote hiring offers access to broader talent pools but requires strong communication and documentation.

7

Non-technical co-founders can be valuable for testing and project management, but require clear communication about timelines.

BUILDING THE MINIMUM VIABLE PRODUCT (V1)

The initial version of a product (V1) should prioritize speed and technical feasibility. For PlanGrid, the V1 involved putting blueprints on an iPad, a seemingly simple task that revealed significant technical challenges with image processing on early devices. Segment's V1 evolved from a failed lecture tool to a core platform feature developed in just one week after identifying a stronger market need. Escher Reality focused on making complex AR algorithms work on mobile devices, iterating rapidly to achieve functional performance. Second Measure spent significant time defining their product offering before development, realizing users needed guided data exploration rather than raw data access.

ITERATION AND USER FEEDBACK

A core theme across all founders is the paramount importance of rapid iteration and continuous user feedback. Segment's journey involved multiple pivots, burning through a seed round before landing on their current product concept, which was validated by quickly launching an open-source library. Escher Reality used developer sign-ups as an indicator of product-market fit. Second Measure initially built a system for investors but quickly learned through user testing that it was too overwhelming, leading them to pivot towards a self-service platform. The consensus is that building in a vacuum is less effective than putting prototypes in front of users early.

BALANCING SPEED WITH ENGINEERING BEST PRACTICES

In the early stages, speed and shipping a functional product often take precedence over extensive testing and scalability. However, this balance shifts as the company grows. PlanGrid acknowledged that some V1 code still exists, representing technical debt. Segment moved from a 'hackathon mode' with no tests to implementing CI/CD and rigorous testing as their team and customer base grew, especially with the advent of enterprise clients. Escher Reality rewrote their codebase when acquired by Niantic to meet stringent quality standards for large-scale games. Second Measure gradually introduced unit tests, CI/CD, and code review processes about 1.5 years post-launch.

CHOOSING TECHNOLOGY AND METHODOLOGIES

The choice of technology for V1 often leans towards what the founding team knows best to enable rapid development, with the understanding that it may be refactored later. Companies like Segment and Second Measure adopted frameworks they were most familiar with, even if not the trendiest. Methodologies like Agile are often adopted loosely ('agile-ish'), with teams incorporating sprints and stand-ups but adapting processes to their specific needs and growth stage. The emphasis is on flexibility and evolving processes as the team expands and complexity increases, with OKRs becoming a framework for goal setting at Segment.

WORKING WITH NON-TECHNICAL CO-FOUNDERS AND DEADLINES

Dealing with non-technical co-founders involves leveraging their strengths, such as early user testing, and managing expectations around deadlines. PlanGrid's co-founder inadvertently provided valuable product testing by breaking the software, highlighting the benefit of external perspectives. Padding estimates and establishing lightweight processes for testing and releases are crucial for managing timelines. While Segment's co-founders were technical, they employed a strategy of asking individuals for their predictions in one-on-ones to get a more realistic sense of upcoming work. The transition for technical founders often involves shifting focus from coding to building an organization.

TEAM STRUCTURE: HIRING AND REMOTE WORK

For core competencies and intellectual property, founders generally advocate for hiring full-time employees rather than relying heavily on contractors, though contractors can serve as a trial for potential hires. PlanGrid and Escher Reality prioritized local, full-time teams initially. Segment found success hiring remote engineers from open-source communities, but later emphasized local hires to foster culture before opening remote positions again. The decision between remote and local hiring depends on the desired company culture and the need to tap into diverse talent pools, with remote work often necessitating better documentation and communication practices. Ultimately, the advice is to hire the best talent, whether local or remote, as the company scales.

Tactical Founder Advice: Key Takeaways

Practical takeaways from this episode

Do This

Prioritize speed and user acquisition in V1, focusing on iteration over perfect engineering.
Engage non-technical co-founders early for testing and feedback.
Pad deadlines, especially considering external factors like app store review times.
Adopt an 'agile-ish' approach, taking elements that work best for your team.
Evolve your processes as the team and company complexity grow.
Consider contractors for non-core or well-defined tasks, or as an interview process.
Focus on building core competencies internally; avoid outsourcing critical IP.
Emphasize clear communication and documentation, especially in remote or larger teams.
Get user feedback constantly and iteratively; build multiple throwaway versions if needed.
Hire a little earlier than you think you need to.
Be adaptable and do what's right for your company size and immediate needs.

Avoid This

Don't over-engineer or focus on perfect scalability/security for V1.
Don't be overly dogmatic about specific methodologies; adapt them.
Don't underestimate the management time required for contractors.
Don't outsource your core technology or intellectual property.
Don't assume initial V1 code will live forever; focus on iteration.
Don't build in a vacuum; get user feedback early and often.

Common Questions

Speed of development and getting users are paramount for V1. It's more important to iterate and incorporate learnings than to have a perfectly robust or scalable product from the start. The goal is to find something people will pay for.

Topics

Mentioned in this video

Companies
Plain Grid

A company that builds software for the construction industry, using AWS, Python, Go, and native development for multiple platforms.

Gainsight

A customer success tool that Segment integrates with.

Second Measure

A company that analyzes credit card data to provide daily insights into company performance, primarily using AWS and columnar data stores.

Blue Apron

A brand that uses Second Measure's data for trend analysis.

Segment

A company offering a single API to collect, organize, and adapt customer data into downstream tools, built on AWS and using Go for its backend.

Asher

Diana's former company, acquired by Niantic, which focused on building backend technology for augmented reality.

Spotify

Mentioned as a company that allows individual engineering teams to self-organize and choose their own development processes.

Apple

The company that announced ARKit, posing a challenge to Asher/Niantic's AR technology development.

GitHub

Used as an analogy for 'the GitHub for construction' by Clean Grid, highlighting version control and collaboration.

Mixpanel

A tool that Segment was initially a competitor to, and which Segment's early internal tool could send data to.

Amplitude

A competitor to Mixpanel, mentioned in the context of Segment's early product development.

Redshift

A columnar data store used by Second Measure on the backend.

Pixar

Ralph Judy's former employer, where he gained graphics experience used in developing the first mobile blueprint viewer.

Niantic

The company that acquired Asher, known for making Pokémon Go, and continuing to build AR backend technology.

Y Combinator

An accelerator program that Segment and Asher/Niantic participated in, influencing their early development and strategy.

Google

The company that developed the OKR model, which Segment uses for goal setting.

Software & Apps
Java

A programming language used for Android native development by Plain Grid.

React

A web development framework used by Plain Grid and Second Measure.

KISSmetrics

A tool that Segment's early internal tool could send data to.

Grails framework

A framework used by Second Measure for their first application, chosen due to the team's existing expertise.

Jira

A project management tool that teams at Segment can choose to use.

Android

An operating system for which native development was done by Plain Grid and cross-compilation was used by Asher/Niantic.

Python

A backend programming language used by Plain Grid.

Objective-C

A programming language used for iOS native development by Plain Grid.

Hacker News

The platform where Segment's co-founder launched their cleaned-up library, leading to significant traction.

Asana

A project management tool that teams at Segment can choose to use.

Windows

An operating system for which Plain Grid developed a full application.

ECS

The container service used by Segment to run its infrastructure on AWS.

Google Cloud

The cloud platform used by Niantic, where Asher's technology was migrated.

ARKit

Apple's augmented reality framework, announced around the time Asher/Niantic joined YC, presenting a competitive challenge.

Go

A backend programming language used by Plain Grid and primarily by Segment.

Kotlin

A mobile development language mentioned as something the speaker loved but might differently approach in future ventures.

Kafka

A messaging system used by Segment for its microservices.

C++

A programming language used for efficiency in developing AR backend technology at Asher/Niantic.

Unity

A development platform mentioned in relation to Asher/Niantic's technology, allowing developers to build AR apps easily.

AWS

The cloud infrastructure provider used by Plain Grid, Segment, and Second Measure.

Google Analytics

An analytics tool that Segment integrates with.

Customer.io

An email tool that Segment integrates with.

lambda

A service leveraged by Second Measure for its data pipeline.

Swift

A mobile development language mentioned as something the speaker loved but might differently approach in future ventures.

More from Y Combinator

View all 187 summaries

Found this useful? Build your knowledge library

Get AI-powered summaries of any YouTube video, podcast, or article in seconds. Save them to your personal pods and access them anytime.

Try Summify free