This chapter covers how a product's value gets stated to the outside world: April Dunford's five-part positioning framework and the workshop mechanics behind it, the internal politics and standing customer-intelligence practices that keep positioning honest, what volatile AI markets broke about positioning execution (including whether to state a point of view on the future and how far ahead to position), and the B2B sales and brand practice that carries the message to buyers. The sales and brand material draws heavily on one company — Wiz, via Raaz Herzberg — with additional AI-era go-to-market detail from Tamar Yehoshua at Glean and a worked case in Intercom's pivot to Fin.
Dunford's definition, from Obviously Awesome (2019), is deliberately absolute: positioning defines how our product is the best in the world at delivering something a well-defined set of customers cares a lot about. The framing matters because it is a test, not a slogan. If you cannot spell that sentence out literally — best in the world, at what, for whom, who cares a lot — the positioning isn't finished, and any messaging decision made downstream of it is being made on sand.
Her own analogy for what positioning does is the opening scene of a movie. An audience arriving with no reference point needs a handful of orienting questions answered before it can settle in: what is this, where does it sit, who should I compare it to, why should I care, is this for me. A customer meeting a product for the first time is in exactly that position, and no feature detail can land before those questions are answered. The implication is that you answer them proactively rather than leaving the customer to infer them — inference is where you lose control of the frame.
Dunford's Five-Part Positioning Framework works in a fixed order: competitive alternatives, distinct capabilities, differentiated value, best-fit customers, and only then market category. The order is the method. Category comes last because it is the frame that makes the already-established value obvious to an already-identified customer; choosing it first means choosing a box and then hunting for reasons you belong in it.
It is worth naming the lineage. Positioning: The Battle for Your Mind (Ries & Trout, 1982) — Ries and Trout, 1982 — coined the term and illustrated it with consumer-market examples, and it is cited here as foundational background, not as a working method: it describes what positioning is without prescribing a repeatable process for producing it. Dunford's framework is presented as the practical successor for B2B tech teams that need to actually produce a positioning statement.
Positioning also has a defined place in the wider strategy stack. Go-to-Market Strategy is one leg of a triad with business strategy and product strategy: business strategy sets market and monetization choices, product strategy decides which problems the product org solves, and go-to-market governs how the resulting product reaches and is sold to customers. A change in business strategy cascades into go-to-market execution and creates the targets product strategy is then built to hit. The ladder above all three — mission, vision, strategy, objectives — is the subject of Strategy, Vision, and the Decision Stack rather than this chapter.
Step one is the hardest and most foundational, and it is not a competitor slide. Competitive-Alternatives Mapping (Status Quo vs. Shortlist) asks what the customer would do if you didn't exist, and separates two buckets: status-quo alternatives (habits, spreadsheets, legacy manual processes, the incumbent tool people default to) and the actual shortlist of named vendors a buyer is weighing. The way to get this is to ask customers and sales directly — "if we didn't exist, what would you do?" — rather than assembling every theoretical competitor the market could produce.
There is a tactical rule attached: don't name an alternative the customer hasn't considered. Doing so isn't a display of confidence, it's a way to stall your own deal — "Whoa, whoa, vibe coding? I wasn't even looking at that. Stop. Oh, I got to go research that." Map to the shortlist the buyer is actually holding.
Step two and three are a whiteboard exercise. Capability-to-Value "So What?" Exercise means listing every distinct capability the product has — often 50 to 100 items for a mature product — and pushing each one through repeated "so what?" questions until it resolves into a customer-facing business outcome: money, time, or risk. The raw list is useless for messaging or sales enablement, so the output is compressed into no more than three value themes, on the premise that nobody can keep more than a handful of value propositions in their heads.
Only then does the customer definition appear. Best-Fit Customer / ICP Reverse-Engineering runs bottom-up: find which existing or target accounts most value those distilled themes, then extract their shared traits — industry, size, maturity, buying trigger — as the ICP. Value themes first, matching accounts second, targeting criteria last. Olsen's Segway case is the cautionary version of getting this backwards: launched at a broad "getting around" market and mostly failed, but adopted by mall cops and tour groups. Working backward from who actually adopted surfaces the real segment and the reason they valued it — people who cover unusually high walking distances — and Olsen offers a single behavioral question, "how many miles do you walk per day?", as a rough proxy for a full segmentation study.
One complementary framework supplies the vocabulary for the actual copy. Jobs to Be Done (JTBD) centers landing-page and product copy on the job the customer is trying to accomplish rather than on a feature list; Intercom applied it to its landing pages roughly fifteen years ago, writing around "resolve a support ticket" instead of enumerating chat-widget features. Where positioning asks who cares and why, JTBD gives you the customer-side words to write it in.
Without a disciplined process, positioning collapses into Vibe Positioning — competing, unresolved opinions between marketing and product. The structural problem is that marketing and product never win that argument, because neither has standing authority to settle it alone. That is why the framework is run as an actual workshop rather than left to drift into personal opinion.
The argument is unwinnable partly because each function is genuinely looking at a different market. Internal Competitor-Definition Divergence Across Functions lays out the four views: product names a horizon competitor — whoever looks most advanced technically, regardless of current customer relevance. Marketing over-indexes on whoever spends most on demand generation, the "scary competitor" defined by share of voice. Sales has arguably the most accurate view of the real shortlist, but systematically undercounts status-quo and no-decision losses, because reps reframe every stall as "That's not a no. That's a not yet, buddy. I'm going to get you next year." And the CEO carries a view frozen at the point they last sold personally and skewed by investor-facing narrative — "Dude, we haven't seen Oracle on our short list in like 5 years. What are you talking about?" No one internally out-argues the CEO, yet that view is the least reliable of the four.
Since no function has the full picture, direct customer data is the tiebreaker. Systematic Customer-Intelligence Practices for Positioning names the standing mechanisms: a customer advisory board convened to settle internal disagreements about what customers actually want; an executive customer-sponsorship program where each executive personally owns roughly four accounts with quarterly check-ins, so leadership's customer knowledge is firsthand and current rather than frozen; win-loss analysis, with win analysis called out as the neglected half — "Win is where we find out what's working"; and a recurring positioning check-in that explicitly solicits input from sales as an early-warning channel for drift. The observation about sales is blunt: in a consumer business, daily access to customer conversations would be prized; in B2B, leadership often skips it and guesses that the product "sucks" instead. "We have that. It's called the sales team. And we never ask them anything." One anecdote makes the sponsorship practice concrete — a multi-quarter relationship with a customer's CIO let the speaker settle an internal PM dispute by calling the actual customer rather than guessing.
Detection and decision are separate steps. Six-Month Positioning Checkpoint Cadence is the detection mechanism: a lightweight "speedrun" on a fixed six-month schedule rather than triggered by a lost deal or a bad quarter, asking three questions — has a competitor changed its claims, has the competitive set caught up on a capability we differentiated on, and has our capability-to-value mapping drifted? The cadence is explicitly a hedge against forecasting fallibility: "We're terrible at predicting the future" and "We have to accept the fact that we are often wrong." COVID is the self-critical example, where some behavior shifts (online shopping) proved durable and others (virtual workshops) fully reverted within a few years.
Repositioning Trigger Criteria then decides what to do with what the checkpoint finds. A genuine reposition is warranted on two conditions only: the value the product actually delivers has changed, or the company has moved into a different competitive category. Absent one of those, drift calls for a refresh, not a reposition — and the distinction is not cosmetic. Dunford's early-career example of moving a product from desktop software to an embeddable mobile database restructured pricing (from a $200 direct download to bulk enterprise sales), changed the sales motion, and required new engineering work (sync capability) to make the product fit its new category. Repositioning is a business-model decision as much as a marketing one.
Dunford's claim, as of a 2026 #mtpcon talk and after applying the process with roughly 60–70 companies through the AI disruption, is that the framework itself hasn't needed to change. What changed is that volatile markets create "a hundred new ways to mess it up." Four New Ways AI-Era Markets Break Positioning Execution concentrates those into four problem areas: needing a point of view about the future; messiness in the competitive-alternative landscape, since the set of "what would the customer do instead?" options is shifting faster than usual; how far ahead to position, described as "the biggie"; and operationalizing customer understanding as the market moves. It is worth saying plainly that items two and four are named but not developed in this chapter's material — the sources go deep on the first and third and leave the others as headings.
Disruptions force repositioning because they change what customers believe is possible, not merely what a product does — COVID, the on-prem-to-cloud shift, and now AI, described as possibly the biggest tech change ever seen. Two AI-specific pressures compound it: SaaS incumbents field customers asking "can't I just vibe code that?", and AI-native companies must work out exactly who or what they are disrupting before they can claim a category.
How Far Ahead to Position (Current vs. Anticipated Competitors) is the sharpest of the four. The rule of thumb: compete against who you actually compete with right now, while building the product for the competitors coming later. Positioning tracks today's shortlist; product strategy tracks tomorrow's. The illustrating anecdote is a February 2025 client at an IT-management company where the CPO believed vibe-coding tools — Replit and Lovable, then about two months old — were already a competitive threat, while the head of sales had never once seen them appear on a real prospect's shortlist. Dunford's assessment was "not yet." Same market, two time horizons, both defensible from where each person sat.
The companion discipline is Position the Current Product, Not the Vision, which separates three things that routinely get conflated: the product vision (the long-term, investor-facing story of where the product is going), the current product (what actually ships, bluntly described as "the piece of junk we actually have right now"), and product strategy (the sequenced steps bridging the two). Marketing and sales messaging get built on the current product against today's real competitors. Product strategy is an internal artifact for explaining the path forward, not a customer-facing positioning input.
The failure mode this prevents is specific to AI-era selling: overhyping a distant AI-native future backfires, because customers respond by delaying purchase of the current, sellable product — "Come back in two years when you got it and I'll buy it. I'm not buying your old crap right now." The fix is not to suppress the vision but to pair it with a staged pathway, which is the subject of the next section. The broader disruption to how product work itself is done sits in AI and the Changing Shape of Product Work.
The counterweight to "position for today" is that in volatile markets customers now need a vendor to state where the whole market is going, not just what the product does. Point of View on the Market's Future as a Positioning Requirement gives the reason: typical B2B buyers don't have the insider exposure to AI discourse that vendors do, and ambiguity is expensive — about half of B2B purchase processes end in "no decision" when the options look similar and the risk of choosing wrong feels high. A stated market point of view gives the buyer a reason to act now rather than wait. Going quiet has a price too: Stack Overflow was widely assumed dead, and Salesforce's stock underperformed despite strong revenue, because markets read silence as irrelevance rather than neutrality. And the obligation is unavoidable in any case — every team that builds a roadmap has already bet on a version of the future and "planted a flag whether you want to or not." Positioning's job is to make that bet explicit.
Not every stated POV is credible. Dunford's test is "why can only we claim this?" If a competitor could paste the same forward-looking claim onto their homepage without anyone blinking, it isn't grounded in differentiated capability. IBM's claimed POV — real-time, open, sovereign AI on mainframes — passes because it's backed by concrete claims a challenger can't simply assert: roughly 70% share of transactional data flowing through IBM systems, standards-body openness, global physical presence. Where the POV comes from matters as much as its content: build it by starting from what the company is genuinely good at and extrapolating how that value persists, rather than reverse-engineering a story that merely sounds futuristic. Usually the assumption already exists inside the company and just needs surfacing; when a team can't write one, that's typically a symptom that the executive team hasn't agreed on its differentiated value, not a writing problem.
Taxonomy of Corporate AI Points of View by Business Model is the evidence that a market POV is less a prediction than a positioning move — each company's stated future conveniently favors the assets it already owns. Anthropic: AI takes over developer tasks. OpenAI: intelligence becomes a metered utility — "I don't care, man. I'm just the electric company. As long as I get my bit, that's good." Replit: agents spontaneously generate bespoke tools, replacing SaaS subscriptions. Microsoft (Nadella): models are commodity, and proprietary context — email, Office, SharePoint — is the moat. Salesforce (Benioff): "our API is the UI," with durable value in the headless back end — "all that data, the logic, the workflows, all these things that are very, very difficult to vibe code." IBM (Krishna): real-time, open, sovereign AI. ServiceNow (McDermott): workflow and governance, not models, are "the whole ballgame." A fourth recurring framing is the sun/planets metaphor — "we are the sun and all the things are planets that revolve around us." Notably, Microsoft's platform-centrism and OpenAI's indifference are treated as compatible rather than opposed: they stake claims on different layers of the same stack, and neither depends on the other being wrong.
The humility clause is explicit. Humans are bad at predicting which changes stick versus revert, and the cautionary examples are executives who staked public identity on speculative futures: Jack Dorsey renaming Square to Block for an all-in blockchain bet, and weeks later Mark Zuckerberg renaming Facebook to Meta for the metaverse — four years on, the metaverse team "barely exists." Treat a point of view as a staked-out position to be revised, not a forecast to be defended.
Meeting Risk-Averse Buyers Where They Are: An AI Maturity Model is the mechanism that lets both things be true at once. Rather than pitching the autonomous end state up front, sequence the sale: map the customer's existing processes; apply agents to low-risk tasks with a human supervising and approving; extend to agent-assisted handling of exceptions as trust builds; then let agents do everything agents are capable of. "You got to show them the pathway to get there that isn't scary." Talk too much about the distant future and present-day deals stall; talk too little and customers assume you have no answer. A staged path holds a credible present-day offer and a credible future POV simultaneously.
Two further pieces of practice attach here. Enterprise AI Determinism Expectation Requires Guardrails, reported by Tamar Yehoshua of Glean, applies the same staging logic to in-product UX: enterprise users intuitively understand a search box but not an open chat interface, and unlike consumers they expect deterministic, repeatable behavior from work tools. The response is guardrails and suggested prompts rather than a blank chat box — the same way search needed autocomplete and refinement suggestions to teach people how to query. And Intercom's Fin: Rebuilding Around AI Anchored to the Customer's Job is the worked case of rebuilding around AI without losing the thread: Fin's positioning stayed anchored to the customer's actual job — resolving a support ticket — with AI as the mechanism rather than the pitch. A support product that showcases AI capability but doesn't resolve tickets faster isn't repositioned, it's decorated.
Positioning is tested at the point where someone decides whether to buy, and the material on that point comes almost entirely from Raaz Herzberg's account of Wiz's early days. The underlying trap is Affirmation Bias vs. Disconfirming-Feedback Bias (Customer Discovery): people default to seeking feedback that confirms what they already believe, and polite affirmation is cheap to produce and comfortable to hear. Wiz — originally founded as "Beyond Networks," a network-security company — was running 10–15 customer calls a day. Weeks in, Herzberg admitted out loud that she still didn't understand what the team was building. That admission, not the volume of calls, is what forced the team to stop counting "sounds interesting" as validation and pivot toward cloud security.
The antidote was a signal taxonomy rather than a mood reading. Generic praise — "this is great," "I love it" — is noise; genuine enthusiasm shows up as unprompted requests for next steps. B2B Buying-Signal Checklist (Pricing, PO Timeline, Internal Introductions) operationalizes that into three concrete asks: does the prospect ask about pricing, about PO and procurement timeline, or for an introduction to the internal team that will actually use and deploy the product? Herzberg treats the third as the strongest, because only real enthusiasm motivates someone to spend their own internal political capital making that introduction. A prospect who stays at the level of one friendly contact, however positive, hasn't shown intent.
High-Effort Ask as a Commitment Filter adds a test you can impose rather than wait for. Wiz sent early proof-of-value candidates a long, detailed technical questionnaire, framed as necessary to scope integration requirements, and used the speed and thoroughness of the response as the qualifying signal. Genuinely motivated prospects returned it quickly and completely; stalls and thin answers revealed interest that wasn't as deep as it sounded — a legitimately useful ask doing double duty as a filter.
All of this rolls up into Pull, Not Push (Sales Philosophy): a sale or proof-of-value engagement should never be pushed onto a customer. The customer's own independent motivation to move the deal forward has to be present, and Herzberg holds that this still applies well past a company's earliest deals. If a deal only advances because the vendor is chasing, that's a signal to withhold further resourcing. "You need that pull from the other end as well, don't push too hard" — the discipline is noticing when you're the one doing all the pushing.
The strongest claim in this chapter's B2B marketing material is Brand-Driven B2B Purchasing: even complex, high-stakes enterprise purchases are made by people responding to brand and story, not by pure spec-sheet comparison. Herzberg's analogy is iPhone versus Android — buyers of consumer and enterprise products alike are swayed by brand affinity, not feature checklists alone. That justifies investing in tone, story, and distinctiveness rather than treating B2B marketing strictly as a features-and-ROI communication exercise.
Security Purchasing as Insurance-Like, Category-Leader-Default Behavior supplies the mechanism for why this bites unusually hard in her category. Security buyers behave like insurance buyers: nobody wants to become the story of "we bought the wrong vendor and got breached," so purchasing defaults to whoever is perceived as the category leader rather than to a rigorous feature-by-feature evaluation. Perceived leadership is self-reinforcing in a way that outweighs marginal product differences — brand changes buyer default behavior, not just buyer preference. She pairs this with a market-size claim: despite how ubiquitous cloud infrastructure feels, an estimated 15–20% of infrastructure has actually moved, making cloud security underpenetrated rather than mature.
Wiz's specific bet was Breaking Category Branding Norms (Wiz's 'Magic' vs. Cybersecurity Fear-Branding) — cybersecurity marketing defaults to dark, fear-based branding, and Wiz went bright and optimistic instead, on the theory that the category's dominant tone was an unexamined convention rather than a requirement. It carried through to physical touchpoints: themed trade-show sets, including a "Wiz of Oz" booth, reportedly drove roughly 5x the traffic of a conventional security-vendor booth on the same budget. Generalized, that's Novelty-Booth Strategy for Low-Awareness Brands: when a brand has almost no awareness in its category, a maximally distinct, talkable booth beats category-standard signaling, because the goal is curiosity-driven foot traffic. "I'll just make it the weirdest booth ever because my goal is just having people look and be like ah what is Wiz." The competition isn't only other cloud-security vendors; it's every other fear-toned booth on the floor.
None of that is recklessness, because Herzberg deliberately runs different risk tolerances in different functions. Asymmetric Risk Postures: Product ('Less Is More') vs. Marketing ('Try Everything') treats product changes as high-cost and semi-irreversible — shipping the wrong thing burns engineering time, accumulates complexity, and is hard to walk back once customers depend on it, so the default is less is more. Marketing experiments are near-zero-cost: a bad campaign, a weird booth, or an off-brand piece of content can be dropped tomorrow, so the default is try everything. Applying one uniform posture across both wastes the asymmetry; risk tolerance should be set by the actual cost of being wrong in that function.
The craft counterpart comes from Tamar Yehoshua, citing Marc Benioff: Treating Marketing Stunts as Product Design describes designing keynotes and campaign stunts — Salesforce's "no software" campaign among them — with the same rigor a product person brings to building a product. The keynote is itself a designed artifact, iterated and judged like a feature. That's a claim about process, not a specific tactic, and it collapses the usual line between product and marketing craft.
Everything above fails if the message itself is unintelligible to someone hearing it cold, and this chapter's last cluster is about that translation problem. Product-to-Marketing Translation Framework (Blind-Spot Reframing) is Herzberg's diagnosis of why product-adjacent people often make poor messagers: deep proximity to a product creates blind spots about how confusing or jargon-laden an explanation sounds from outside. Internally, an answer that's "close enough" or slightly fuzzy passes because everyone shares context; a buyer hearing it cold has no such scaffolding. Marketing's job in this framing is a translation function — converting internally-acceptable, technically-precise-but-fuzzy answers into crystal-clear, jargon-free external statements. That gap is part of what pulled her from product management into marketing at Wiz.
The operational rule it produces is The 'Dummy Explanation' (Assume Zero Prior Knowledge): when explaining the company or product externally, assume the audience knows nothing — not the category, not the company name, not the problem it solves. Wiz uses this as a named internal principle. It matters most for a category with real conceptual overhead like cloud security, and for founders who assume shared context because they've explained the thing hundreds of times internally.
Which sets up the trap in "The Bubble": Internal Message Fatigue vs. Customer Freshness. The team's boredom with a message is not evidence that customers are tired of it. Employees who live with the positioning daily hit repetition fatigue long before any individual customer does — most customers are on their first or second exposure when the internal team has seen it hundreds of times. Herzberg frames the discipline as stepping outside "the bubble": don't retire external messaging because it feels stale from the inside; check for actual external fatigue — declining engagement, direct customer feedback — first. Assume-nothing explanations specifically need to be repeated past the point where they feel obvious to the team.
A final heuristic runs across both product and marketing decisions. "Keep It Simple": Complexity as a Signal You Haven't Found the Right Answer Yet treats complexity as diagnostic: if a proposed feature, answer, or decision feels overly complicated to explain or execute, that complexity is itself evidence the current solution isn't the right one — not an obstacle to push through with more effort. The prescribed response is to stop forcing the complicated version, set it aside, and come back later, on the assumption that the right answer will feel simple once found.
One honest note on coverage. This chapter is deep on positioning as a discipline and on brand and early-sales instinct, but the go-to-market leg is defined more than it is worked out — pricing, channel strategy, demand generation, and sales-motion mechanics barely appear beyond the pricing changes named as part of a genuine reposition. The B2B sales material is drawn almost entirely from one company at one moment in its growth, and two of Dunford's four AI-era failure modes are named without being developed. Read the sales and brand sections as one well-documented account rather than as a general theory.