samarv The AMPED Lifecycle: Balancing Speed, Quality, and Value
The AMPED (Analytics, Marketing, Product, Engineering, Design) lifecycle is a system designed to ensure that product development is a "virtuous cycle" of improvement. It moves beyond simple "shipping" to track exactly how long value takes to reach the customer and whether that value meets a binary standard of excellence.
The P-Gate Workflow
Categorize every project as Small (under 1 month), Medium, or Large. Track the time spent in each of the following phases:
- P-Strat (Strategy): The initial concept pitch. Align the project with the annual strategy white paper and define the "Chess Move" (how this makes the product better relative to competitors).
- P0 (Problem Definition): A one-page spec focused strictly on the user pain point and the metric to be moved.
- P1 (Solution Definition): The technical and design specification of the product to be built.
- P2 (Value Validation): Post-ship analysis. This phase is only complete when the original target metric from P0 is measured and moved.
Operationalizing Velocity and Quality
1. Velocity Benchmarking (The "Golf" Method)
Treat velocity as a game of golf—compete only against your own past performance.
- Collect Cycle Data: Record the duration of each P-gate across all 50+ feature teams.
- Calculate Medians: Establish the median time for a "Small" project to move from P0 to P1.
- Identify Roadblocks: If a team exceeds the median, leadership should not penalize them but instead "raise a hand" to remove technical or organizational blockers (e.g., dependency issues, lack of design resources).
2. Binary Quality Triage
Avoid long, subjective documents defining "quality." Use a classification system instead.
- Monthly Review: Design leadership reviews every feature shipped in the last 30 days.
- Binary Labeling: Categorize each ship as either High Quality or Not High Quality.
- Color Comparison: Use these examples to train the org. Just as you learn "Pink vs. Red" by seeing them, PMs learn quality by seeing "High vs. Low" examples.
- Feedback Loop: Share the "Not High Quality" list with teams to build "reinforcement learning" for future sprints.
3. The 80/50 Roadmap Rule
Maintain a rolling 6-month roadmap updated every 3 months to balance enterprise needs with agility:
- Months 1–3: Commit to an 80% confidence level. These items are nearly certain to ship.
- Months 4–6: Commit to a 50% confidence level. This allows the team to pivot based on new competitive threats or "hitting a brick wall" (learning a path is wrong) without breaking promises.
Examples
Example 1: Diagnosing a Slow Feature Launch
- Context: A team is building an asynchronous video tool (e.g., Talktrack).
- Input: The project is tagged as "Medium" and is currently in the P1 phase.
- Application: The PM notices the P0-to-P1 time has hit 45 days, while the org median for Medium projects is 20 days.
- Output: The PM raises a hand in the Monday leadership meeting. They identify a "blocker" in engineering architecture. Leadership reallocates a senior engineer for 48 hours to resolve the bottleneck.
Example 2: Calibrating Design Standards
- Context: A new "Sharing" flow is released.
- Input: The design leadership triage.
- Application: The flow is labeled "Not High Quality" because the button placement is inconsistent with the platform's core UI.
- Output: The example is added to the "AMPED Quality Gallery." The feature team sees the specific comparison and adjusts the next iteration to meet the "High Quality" benchmark.
Common Pitfalls
- Focusing on Speed over Value: Shipping fast doesn't matter if the P2 phase shows the metric didn't move. Never close a project at "Ship"—only at "Impact."
- Over-Defining Quality: Don't write 20-page "Quality Guidelines." People won't read them. Show them 10 examples of "High Quality" and 10 examples of "Not High Quality."
- Siloing Marketing: If Marketing isn't in the "AMPED" loop from P-strat, the product might be functional but fail to "capture the imagination" of the market because the positioning is an afterthought.
- Competing Between Teams: Using velocity data to rank teams against each other. Use data only to find where leadership needs to help unblock a team.