Inside Gong: How teams work with design partners, their pod structure, autonomy, trust, and more | Eilon Reshef (co-founder and CPO)
Elan Reshev, co-founder and CPO/CTO at Gong, shares secrets to their success, including their "pod model" for product teams, working with 6-12 design partners per feature, fostering autonomy, making quick decisions, and building AI products.
Deep Dive Analysis
14 Topic Outline
Gong's Unique Pod Model for Product Teams
Extreme Collaboration with Design Partners
Finding and Coordinating Design Partners Effectively
Balancing Customer Feedback and Product Vision
High Success Rate of Gong's New Features
Cultivating Autonomy and Trust in Teams
Implementing a Culture of Autonomy and Trust
Philosophy on Speed and Decision-Making
Lessons from Early AI Adoption and Development
Building Effective AI Teams and Measurement
The Spiral Method for Rapid Learning
Power of Narrowing the Initial Customer Profile
Past Business Failure and Key Learnings
Lightning Round: Books, TV, Products, and Mottos
4 Key Concepts
Pod Model
Gong's product team structure where each pod is a cross-functional unit consisting of a product manager, UX designer, fractional writer, fractional analyst, engineering team leader, and 5-7 engineers. These pods are autonomous and responsible for specific 'jobs to be done' or product outcomes.
Design Partner Approach
An extreme form of customer collaboration where each product pod works directly with 5 to 24 design partners, showing them half-built features and iterating based on their real-time feedback. This method significantly increases the utility and adoption rate of new products and features.
Spiral Method for Learning
A technique for quickly learning complex topics by starting with basic questions, talking to multiple experts, and continuously asking 'who else should I speak with?' The process spirals deeper into the topic until new information becomes scarce, indicating a sufficient level of understanding for one's role.
Hanlon's Razor
A philosophical principle stating, 'Never attribute to malice that which is adequately explained by stupidity.' It suggests that when someone's actions cause a negative outcome, it's often due to lack of knowledge, oversight, or incompetence rather than intentional harm, promoting a more forgiving interpretation of behavior.
9 Questions Answered
Gong organizes its product teams into autonomous pods, each comprising a product manager, UX designer, fractional writer, fractional analyst, engineering team leader, and 5-7 engineers, responsible for specific product outcomes or 'jobs to be done'.
Each Gong product pod works directly with 5 to 24 design partners, showing them partially built features and iterating rapidly based on their feedback. This hands-on approach ensures new products and features address real customer needs.
Product managers are expected to discern 'must-have' requests from unique ones, often by asking customers about their current satisfaction and aiming to improve it significantly. While specific customer needs are considered for large deals, the design partner program focuses on building solutions applicable across the broader customer base.
Over 95% of the capabilities Gong builds are used in a significant way by customers, largely due to the extensive direct collaboration with design partners.
Eilon Reshef believes that giving teams autonomy allows individuals to bring their full selves and motivations to their work, leading to better short-term and long-term results, higher velocity, and increased engagement.
Leaders must be willing to 'let go' of control, accept potentially more mistakes, and acknowledge a trade-off in visibility. Peers (e.g., sales, CFO) also need to trust the process, and teams are expected to proactively solicit feedback and drive their own reviews.
People should not assume LLMs solve everything; specialized AI expertise, core competencies, and robust measurement systems (like the Elo system for quality) are still crucial for advancing beyond initial versions and achieving operational rigor.
Using the 'spiral method,' one starts by asking basic questions, then speaks to multiple experts, continuously asking for more contacts. This process deepens understanding until new information becomes scarce, indicating sufficient knowledge for one's role.
Starting with a very narrow ICP allows a company to create a 'small pond' where customers talk to each other, fostering a viral effect and enabling the development of a highly focused, repeatable product-market fit, as seen with Gong's initial targeting.
11 Actionable Insights
1. Work with Many Design Partners
Have every product pod work hand-in-hand with 6 to 12 (sometimes up to two dozen) design partners on new products and features. This approach leads to a near 100% success rate for features being used by a significant number of people, as customers know what they need.
2. Empower Teams with Autonomy
Give teams significant autonomy and trust, allowing them to bring their full selves to their work and decide how to progress on their agenda. This approach leads to much better short-term and long-term results, higher motivation, and engagement.
3. Make Quick, Decisive Choices
Optimize for making decisions quickly, even large “one-way door” decisions, rather than overthinking or waiting for all information. Many decisions are 51/49, so making a choice and moving forward often doesn’t lead to radically different outcomes, saving time and mental energy.
4. Retain In-House AI Expertise
Don’t assume LLMs solve everything; retain core AI competencies and expertise within the company (e.g., data scientists, prompt engineers). This ensures understanding of what’s doable, proper measurement (e.g., using Elo system), and operational rigor to advance beyond V1.
5. Replicate Cross-Functional Pods
Organize product teams into “pods” consisting of a product manager, UX designer, fractional writer/analyst, engineering team leader, and 5-7 engineers, giving them autonomy to solve problems and work with customers. This structure helps in scaling and ensures reasonable direction.
6. Hire a Research Coordinator
Employ a dedicated research coordinator to manage the logistics of connecting product managers with design partners. This person handles outreach, identifies target customer profiles (ICP), and schedules meetings, freeing PMs from coordination stress.
7. Interpret Customer Feedback Wisely
Product managers should develop the core skill of discerning “must-have” requests from “nice-to-have” feedback, aiming to move customers from a 6 to an 8 or 9 in satisfaction. This ensures products are built for the broader customer base, not just one-off requests.
8. Trade Visibility for Velocity
As a leader, be willing to give up some visibility and control over every feature to empower teams with more autonomy. This trade-off can lead to higher velocity, increased morale, and ultimately better products.
9. Use Spiral Method to Learn
To quickly learn complex topics, use the “spiral method”: talk to one person, then ask them who else to speak with, and continue spiraling through conversations until you hear nothing new at your desired level of understanding. This helps confirm learning and efficiently gather diverse perspectives.
10. Start with a Narrow ICP
Begin by targeting a very narrow Ideal Customer Profile (ICP) with specific constraints (e.g., selling in the US, English, over video, specific software value). This creates a “small pond” where customers talk to each other, fostering viral effects and making it easier to achieve product-market fit and scale.
11. Apply Handel’s Razor
Adopt “Handel’s razor”: “Never attribute to malice that which is adequately explained by stupidity.” This mindset helps avoid misinterpreting others’ behavior as malicious, instead assuming they “didn’t know, didn’t care, weren’t trained,” which can improve interactions and understanding.
5 Key Quotes
I would say very close to 100% of the features we build end up being used by a significant number of people.
Eilon Reshef
I hate terms such as risk because that's a very ambiguous term, but just the risk of building something, you're not going to know if it's going to get used.
Eilon Reshef
I do believe customers know much better than what they need. And then myself or my colleagues in the executive team, whoever else it's, you talk to a customer and they kind of describe the pain. They might not know how to build it or what's the right way to implement it, but the pain should be there.
Eilon Reshef
It's a very selfish thing. It's a very personal thing. I just think you get more from everybody if you kind of let them be themselves and do things in the way that they believe is the right way of force within limits.
Eilon Reshef
Never attribute to malice, that which is adequately explained by stupidity.
Eilon Reshef
3 Protocols
Gong's Pod Structure
Eilon Reshef- Assign one Product Manager.
- Assign one User Experience Designer.
- Assign fractional Writing and Analyst support.
- Assign one Team Leader from Engineering.
- Assign 5-7 Engineers (backend and frontend).
- Give the pod an agenda (e.g., a specific 'job to be done' like a forecast product or sales engagement).
- Grant the pod autonomy to identify how to solve problems and work with customers.
- Have the pod work with 5-24 design partners directly.
- Measure progress and iterate based on design partner feedback.
Design Partner Coordination (Gong's Research Coordinator Role)
Eilon Reshef- Product Managers (PMs) communicate their target market (ICP) and what they want to learn from customers to the Research Coordinator.
- The Research Coordinator uses a micro CRM to sift through Gong's customer base, slicing and dicing it based on PMs' criteria (e.g., head of RevOps at mid-size companies, IC seller at enterprise).
- The Research Coordinator runs micro email campaigns to identified customers.
- The Research Coordinator sets up meetings in the PMs' calendars with potential design partners.
- Double-check with customer success to avoid contacting frustrated customers or those in negotiation (optional but recommended).
The Spiral Method for Learning Complex Topics
Eilon Reshef- Identify a complex topic you want to learn (e.g., deep learning).
- Find an initial person to talk to and ask basic questions about the topic.
- After the conversation, ask 'Who else should I be speaking with?'
- Speak with the recommended individuals, gathering more perspectives and information.
- Continue this process, spiraling through conversations, until the amount of new information you're hearing significantly decreases (e.g., new stuff is 10% or 5% or 0%).
- Recognize that you have reached a sufficient level of understanding for your role, even if not a specialist.