Marty Cagan · Oct 21, 2026
Silicon Valley Product Group argues experimentation should shape what product teams build
Markdown · JSON · Original source
’s article argues that as AI makes building and deploying product capabilities faster and cheaper, deciding what to build becomes more important. It distinguishes the project model, in which teams deliver assigned projects and features, from the product model, in which teams solve problems and are responsible for outcomes. The article says experimentation should extend beyond optimization—testing ways to improve an existing product—to innovation: learning what to build and why. It describes qualitative experiments, such as testing prototypes with users, and quantitative experiments, which use data to assess outcomes, as complementary ways to learn. AI can help with tasks such as analyzing user data and evaluating tests, the article says, but experiments still inform decisions and strategy. For AI products, the article notes that evaluations, or , add a layer to product discovery; it says quantitative tests can be more expensive, take longer, and yield less durable conclusions. It advocates a hybrid model: product teams own experiments with coaching and assistance from experimentation professionals. The proposed shift is from counting tests to assessing business and customer impact. The article does not offer evidence for its claims about high failure rates or show that the proposed organizational approach produces better outcomes. What results
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Questions using this source
These authored EM reads cite this source. Follow their assumptions and limitations; inclusion does not imply agreement.
Original source
· Article
Source details
Original source date: Oct 21, 2026 (2026-10-21T15:53:41+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-05T03:53:08.444Z. First observed: 2026-11-04T12:33:18.400Z. Retrieved: 2026-11-05T03:52:28.876Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The article asserts that quickly built and released products have had high failure rates and serious consequences, but supplies no figures or supporting cases in the provided text. The article advocates the hybrid model but provides no comparative results showing whether it improves business or customer outcomes. The author is not identified in the provided article text.
Open question
What results would show whether the hybrid model improves business and customer outcomes compared with other approaches?
Source-stated
The article argues that AI’s lower building and deployment costs make choosing what to build a more critical product skill.
As AI makes the actual building and deployment of product capabilities faster and cheaper, everything depends on what we choose to build.
Source-stated
The article distinguishes the project model, which assigns teams projects and features, from the product model, which gives teams problems to solve and responsibility for outcomes.
In contrast, in the product model, product teams are instead given problems to solve, and they are then empowered to come up with a solution that delivers the necessary business results (outcomes).
Source-stated
The article argues experimentation should contribute to innovation and product discovery, not just optimization of existing products.
It means expanding experimentation beyond optimization to innovation, and beyond proving outcomes, to influencing outcomes.
Source-stated
The article advocates a hybrid model in which product teams own experiments and experimentation professionals coach and assist them.
It lets the product teams own the experimentation, and take responsibility for outcomes, yet with the assistance and coaching of experimentation professionals.
Source-stated
The article says quantitative tests for AI products may be costlier, slower, and less durable as a basis for conclusions.
Product discovery includes the eval process, and many of the judgements about the product are subjective. Moreover, an AI product changes the calculus for the use of quantitative tests, as they are generally more expensive and take longer, and the conclusions may not be as durable.
Source and contributor context
Marty Cagan · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Evals
Structured tests of what an AI system can do or how it behaves, run before or after release.
Silicon Valley Product Group
Silicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Editorial provenance
Version 1. Publication ID: 264aeac9-d630-436f-8f7a-9c138a120029.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T03:53:08.444Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: fe921170edb21268e8a724b22bcf59f7cc7f7c7923a78dd589d17b85b72ba33c. Source text hash: cbb5b07cadd4da648e009e256700cd04ac23f30a4df84bbf34dd539ed38ba011.
Revision and source-check history
Version 1 · 2026-11-05T03:53:08.444Z
Silicon Valley Product Group argues experimentation should shape what product teams build
’s article argues that as AI makes building and deploying product capabilities faster and cheaper, deciding what to build becomes more important. It distinguishes the project model, in which teams deliver assigned projects and features, from the product model, in which teams solve problems and are responsible for outcomes. The article says experimentation should extend beyond optimization—testing ways to improve an existing product—to innovation: learning what to build and why. It describes qualitative experiments, such as testing prototypes with users, and quantitative experiments, which use data to assess outcomes, as complementary ways to learn. AI can help with tasks such as analyzing user data and evaluating tests, the article says, but experiments still inform decisions and strategy. For AI products, the article notes that evaluations, or , add a layer to product discovery; it says quantitative tests can be more expensive, take longer, and yield less durable conclusions. It advocates a hybrid model: product teams own experiments with coaching and assistance from experimentation professionals. The proposed shift is from counting tests to assessing business and customer impact. The article does not offer evidence for its claims about high failure rates or show that the proposed organizational approach produces better outcomes. What results
Questions using this source
These authored EM reads cite this source. Follow their assumptions and limitations; inclusion does not imply agreement.
Original source
· Article
Source details
Original source date: Oct 21, 2026 (2026-10-21T15:53:41+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-05T03:53:08.444Z. First observed: 2026-11-04T12:33:18.400Z. Retrieved: 2026-11-05T03:52:28.876Z.
Limits
The article asserts that quickly built and released products have had high failure rates and serious consequences, but supplies no figures or supporting cases in the provided text. The article advocates the hybrid model but provides no comparative results showing whether it improves business or customer outcomes. The author is not identified in the provided article text.
Open question
What results would show whether the hybrid model improves business and customer outcomes compared with other approaches?
Source-stated
The article argues that AI’s lower building and deployment costs make choosing what to build a more critical product skill.
As AI makes the actual building and deployment of product capabilities faster and cheaper, everything depends on what we choose to build.
Source-stated
The article distinguishes the project model, which assigns teams projects and features, from the product model, which gives teams problems to solve and responsibility for outcomes.
In contrast, in the product model, product teams are instead given problems to solve, and they are then empowered to come up with a solution that delivers the necessary business results (outcomes).
Source-stated
The article argues experimentation should contribute to innovation and product discovery, not just optimization of existing products.
It means expanding experimentation beyond optimization to innovation, and beyond proving outcomes, to influencing outcomes.
Source-stated
The article advocates a hybrid model in which product teams own experiments and experimentation professionals coach and assist them.
It lets the product teams own the experimentation, and take responsibility for outcomes, yet with the assistance and coaching of experimentation professionals.
Source-stated
The article says quantitative tests for AI products may be costlier, slower, and less durable as a basis for conclusions.
Product discovery includes the eval process, and many of the judgements about the product are subjective. Moreover, an AI product changes the calculus for the use of quantitative tests, as they are generally more expensive and take longer, and the conclusions may not be as durable.
Source and contributor context
Marty Cagan · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Evals
Structured tests of what an AI system can do or how it behaves, run before or after release.
Silicon Valley Product Group
Silicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Editorial provenance
Version 1. Publication ID: 264aeac9-d630-436f-8f7a-9c138a120029.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T03:53:08.444Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: fe921170edb21268e8a724b22bcf59f7cc7f7c7923a78dd589d17b85b72ba33c. Source text hash: cbb5b07cadd4da648e009e256700cd04ac23f30a4df84bbf34dd539ed38ba011.
uncertain · 2026-11-05T03:53:08.444Z · anthropic/claude-opus-5.5
Publication: 264aeac9-d630-436f-8f7a-9c138a120029
attribution — pass: The draft attributes the argument to Silicon Valley Product Group as publisher, consistently uses 'the article argues/says', and notes that the author is not named in the text. No source assertion is presented as established fact.
This talk is focused on the product organization, the people responsible for building the products the company sells.
summary — uncertain: The content faithfully covers the AI-lowers-build-cost thesis, the project vs product model, the shift from optimization to innovation, qualitative and quantitative experiments, evals, the hybrid model, and the shift from counting tests to measuring impact. However, the summary ends with a dangling fragment, 'What results', which appears to be a truncated copy of the question. This leaves the reading malformed. Smallest repair: delete the trailing fragment. The title and question otherwise follow from the source.
This is why the hybrid model is often the most effective, and what we generally advocate.
date — pass: originalPublished 2026-10-21T15:53:41Z comes from the publisher page and matches the feed entry. Nothing in the text points to an earlier publication of this narrative version. The keynote delivery date is not given and does not affect the publication date.
support — pass: Each claim is matched by an exact excerpt. The draft's 'can be more expensive' and 'less durable' wording stays close to the source's 'generally' and 'may not be as durable' hedging. The limitations correctly note that the high-failure-rate assertion and the hybrid model's effectiveness are unsupported by data in the text. The question is open and not leading.
Initially companies thought they could simply build quickly, release into the wild, and see what sticks, but the high failure rates come with serious consequences
media — pass: The selected media is the verified direct link to the original SVPG article. It is the most direct way to reach the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Silicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Explore Silicon Valley Product GroupStructured tests of what an AI system can do or how it behaves, run before or after release.
Definition referenceSilicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Explore Silicon Valley Product GroupSilicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Explore Silicon Valley Product GroupStructured tests of what an AI system can do or how it behaves, run before or after release.
Definition referenceSilicon Valley Product Group, founded by Marty Cagan, publishes essays and coaching material on product management and product leadership.
Explore Silicon Valley Product Group