framework · strategy

7Ps marketing framework for PMs

Best for: GTM, pricing, market entry, and product launch questions where the service delivery layer matters. Best for B2B SaaS, AI products, and marketplace contexts where how the product gets delivered is as important as what it does.

Updated Jun 2026 Calibrated to the strong-hire bar

The 7Ps exist because McCarthy’s original 4Ps (Product, Price, Place, Promotion) could not handle service businesses. Booms and Bitner added People, Process, and Physical Evidence in 1981, specifically to capture what distinguishes one airline, bank, or consulting firm from another when the “product” is largely the same. In a PM interview, the 7Ps appear most in strategy questions: GTM, pricing, market entry, competitive response, and product launch. The signal the interviewer is reading is not whether you can name all seven. It is whether you know which three are contested for this product, at this moment, and can reason through them with specificity.

The seven Ps and what they mean in practice

Product: What problem you are solving and for which segment, not a feature list. In 2026, feasibility is largely table-stakes. You can build almost anything. That means this P is no longer about “can we build it” but “is this the right problem, and will people find it lovable enough to pay?” Notion’s product is not a note-taking app. It is a workspace that replaces the proliferation of tools a small team accumulates before they can afford an operations hire.

Price: The PM’s most direct test of viability. AI products have made this P more complex, not simpler. You now have four credible models: per-seat, usage-based (token or API call), outcome-based (pay-on-success), and platform-plus-consumption. Stripe’s pricing for its core Payments API (2.9% + 30 cents per transaction) is a usage-based model where pricing scales with customer success. Choosing the wrong model is a GTM error, not just a finance question: a seat-based model applied to a product with uneven usage patterns will stall land-and-expand. A usage-based model applied to a product where costs are unpredictable will crater enterprise procurement.

Place: Where and how the product reaches users. In 2026, Place includes PLG loops, API integrations, agent-to-agent marketplaces, AI discovery surfaces (Perplexity, ChatGPT, Google AI Mode), and distribution partnerships alongside traditional channels. Shopify’s early Place strategy was developer marketplaces and agency partnerships, not paid acquisition. A GTM answer that only names the app store and social ads is missing the channels that now drive the majority of B2B inbound at launch.

Promotion: The primary message tied to the job-to-be-done, delivered through the right channel. PMs co-own the positioning brief but promotion execution belongs to PMM and marketing. In an interview, your job is to demonstrate you can brief a campaign, not write copy.

People: Who delivers the product experience. This P is bifurcated in 2026. One side is the human team: customer success, implementation, support, and the sales motion for enterprise. The other side is the AI or agent layer users interact with directly. For a product like Glean or Sierra, the “person” a user talks to is largely a model. The quality, accuracy, and recovery behavior of that model is a People decision as much as a hiring one.

Process: How the service is delivered, step by step. For AI products, this is the hardest P to get right. The process IS the model’s behavior: what happens when it produces a wrong output, how users correct it, what guardrails exist, and how the human-in-the-loop is designed (or not). A PM who describes an agentic product’s process as “the AI handles it” has not answered the question. Non-determinism is a process design problem.

Physical Evidence: The tangible proof that the product works, before the user has experienced it. For digital products, this translates directly to trust signals: security certifications (SOC 2, ISO 27001), case studies with named customers and specific metrics, G2 or Capterra review counts, published eval benchmarks, sandbox or trial environments, uptime pages, and SLA terms. This is the P PMs most frequently overlook. At Stripe, the documentation quality is Physical Evidence. At Notion, the 30-day template gallery is Physical Evidence. For a new AI vendor competing for an enterprise security contract, a published accuracy report and a named reference customer are Physical Evidence. Without these, the pipeline stalls at security review, not at the product demo.

Who owns what

Interviewers probe this because candidates often talk as if they own all seven. Be accurate.

  • PM owns: Product (problem selection and scope) and Price (model, packaging, and unit economics).
  • PM co-owns: Place (channel strategy, especially PLG mechanisms and integration partnerships), Process (delivery design, particularly for agentic products), and Physical Evidence (trust infrastructure: what gets built vs what marketing writes about it).
  • PMM and marketing own: Promotion. PM provides the positioning brief and the segment/job-to-be-done framing.
  • Shared with HR, engineering, and ops: People. PM sets the bar on model quality and human support design; they do not own recruiting or CS team structure.

Strong and weak answers

strong

"You asked how I'd GTM a new AI writing tool targeting enterprise. I'd focus on three Ps where the strategic tension is real: Price, Process, and Physical Evidence. Price first: enterprise buyers default to seat-based contracts because procurement understands them, but our usage pattern is uneven: a senior writer might generate 50 drafts a day, a junior might generate two. A platform fee plus a usage tier above a threshold gives us predictable revenue for finance and aligns cost with value for buyers. That model also makes land-and-expand easier than a seat count debate every renewal. Process second: our product delivers value through an AI layer and every enterprise buyer will ask 'what happens when it gets the tone wrong or fabricates a citation?' I'd design a human-review step into the default workflow, not as a workaround but as a named feature. The process question is where we win or lose the security review. Physical Evidence third: we have no brand recognition. Before we open the waitlist, I'd build a sandbox environment with sample briefs, publish an eval showing accuracy on a standard benchmark, and secure three design-partner case studies with named companies and specific metrics (not 'reduced time' but '40% fewer revision cycles on regulatory filings'). Promotion I'd deprioritize early. A PLG motion through the free tier gets us usage data and real-world case studies faster than paid acquisition, and those case studies become the Physical Evidence for the next wave of buyers. The metric that tells me the GTM is working: free-to-paid conversion among accounts that complete three full drafts in the first 14 days, not raw signups."

weak

"The 7Ps are Product, Price, Place, Promotion, People, Process, and Physical Evidence. For an AI writing tool, the product is the writing tool, the price would be competitive, we'd promote it on LinkedIn and Product Hunt, and we'd make sure the process is smooth for users." This is recitation. Equal time on all seven signals no prioritization. "Competitive" is not a price. "Smooth" is not a process design. Physical Evidence (the P that most often kills enterprise deals) got five words. The interviewer cannot tell whether this candidate has shipped anything or read a textbook.

Use it, do not recite it

The 7Ps is a checklist for organizing your thinking before you speak, not a structure you walk through in order. Strong candidates name the two or three Ps where the real risk lives for this specific product, make a recommendation with trade-offs, and explicitly note what they would deprioritize and why. Saying “I’d spend less time on Promotion early because PLG distributes faster than paid for this segment” is more impressive than a complete tour of all seven.

The LeadingTheProduct reframe reorders the Ps as Problem, Purpose, Position, Performance, Price, Promotion, and Practice. A useful alternative when the question is about product strategy rather than GTM. Know both, and pick the one that fits the question.

The 2026 reframe

Feasibility is free, which collapses the traditional Product P from “can we build this” to “is this the right problem and will anyone love it enough to pay.” That shifts the center of gravity toward the Ps that are actually contested.

Price is harder than it has ever been. AI products can be priced on usage, seats, outcomes, or value-share. The right model often determines adoption trajectory more than the feature set. A PM who can only say “freemium with an enterprise tier” has not cleared the bar.

Process now includes the AI itself. For agentic products, the process is the model’s behavior: its accuracy, its recovery paths, and its guardrails. These are product decisions, not engineering footnotes.

Physical Evidence determines whether you get to the demo. In a market where every vendor claims AI-powered everything, the trust signals you build before a buyer commits (evals, named case studies, sandbox environments, transparent error rates) are the difference between a qualified pipeline and a lot of demos that go nowhere.

The Ps that 2026 PMs directly control: Product (problem selection and experience quality), Price (model and packaging), Process (agentic reliability and recovery), and Physical Evidence (trust infrastructure). Promotion and Place increasingly belong to PLG loops and distribution partnerships. See feasibility is free for the broader context, and proving viability for how to pressure-test the Price P in an interview.