A Viewpoint from Bold to Win
SaaS products are strange beasts. You win buyers long before you earn users from a resulting deal. Buyers who buy and users who use are two different customers and you have to win both. Sure, you can dodge the buyer for a while by going direct-to-end-user. But eventually, as Slack and Zoom both learned, to scale in SaaS you win logos one deal at a time. The buyer is a customer, and it’s not a one time event. You engage with the buyer at every renewal — and simultaneously the buyer is quietly monitoring the market for alternatives, from other software vendors eager to take your use case as well as emerging competitors fighting for your customer;s business.
So in SaaS bold product leaders have to obsess over how the product is bought, not just how it’s used. As Clayton Christensen highlighted in The Innovator’s Dilemma your company is effectively owned by your current users. Your feedback channels are all tuned to listen to this audience. You have an abundance of features, bugs, and issues and research from current users — far less from current buyers, and almost nothing from future buyers. So over time, the gravitational pull of your product, over-index on the customers you already have and what they wanted yesterday—not how they buy tomorrow.
And that’s a problem because the market and the competition decide whether your customers stay — and whether their spend grows — not your roadmap. You can ship exactly what your customers told you they wanted, line for line, and still miss a market inflection point and watch them leave anyway. OpenAI chased the consumer software and ad business, and watched Claude catch up and take the enterprise market out from under it. Nobody’s roadmap failed. The market moved, and OpenAI’s roadmap was pointed at what consumers wanted, and not the enterprise.
I started seeing this pattern over and over and I needed a framework — one that stays fixed on how the product competes in the marketplace, and then organizes the product to deliver against that. I had experience with a foundation in a framework called XO, and paired it with an idea I’ve been working with for years: increasing returns.
The foundation: Experience Outcomes
At Microsoft Office I became acquainted with Experience Outcomes — XOs — a framework from Steven Herbst, who also gave us “UI Traps.” Herbst developed XO inside Microsoft and carried it to Amazon; I carried it to Nordstrom and into the work I’ve done since. It has shaped how I read research for more than a decade, and Winning Product Experiences and Eroding Product Experiences are direct descendants of it. Credit where it’s due.
The XO was a deceptively simple idea, which is exactly what made it a breakthrough. A product is the totality of the jobs it helps people do — and a job is a what, not a how (“get to known content before we argue about what to watch,” not “open the search menu”). An XO took a job and bolted measurable success criteria onto it: a direction (are we moving the metric up or down), a unit of measure (how we know we’re done), and an outcome (the value the customer gains). “Get to a known piece of content in under thirty seconds” is an XO. A feature is not.
What made XO powerful was discipline. It refused to confuse whats with hows. It held teams accountable for the unglamorous, low-status jobs that decide whether an experience actually feels good — not just the technically exciting ones that attract volunteers. And it gave the organization a single currency of quality: a set of XOs became a testable thesis for success — satisfy the most important ones better than the competition, and you win. And you could get customer feedback on each importance.
I still believe that, and it worked phenomenally when I was focused on the end user in consumer tech. But as I dug deeper into SaaS it didn’t translate the way I wanted. XO measures the job to be done. It does not, on its own, measure the jobs you win and the jobs you lose. Improve an XO in e-commerce and revenue goes up. Improve an XO in SaaS and you may be polishing a job your customer’s value but that no longer close deals.
The catalyst: increasing returns
Years ago I wrote a piece called Prioritize Roadmaps for Type 2 Impact (Medium). The argument: most product teams prioritize for diminishing returns without ever naming it. When someone says “let’s lift our onboarding completion rate” or “optimize the conversion funnel,” they’re saying — of the customers who already reached this step, let’s convert more of them. The ceiling is 100%. Every improvement creates less upside than the last, while the difficulty of the next improvement goes up.
Increasing returns is the opposite, and it has no ceiling. You invest in the things that pull more customers toward the product to begin with — the things that strengthen the offer itself, so its gravity grows. The rich get richer. The QWERTY keyboard didn’t win because it was good; it won because each new typist who learned it made it more worth the next person’s while to learn it too. There is no maximum pull a product can have. Just imagine the market for bottled water—in the 60s American’s only consumed a gallon and a half of water per year and now that number is more than 40 gallons per person (source: USDA).
Here’s the test, stated plainly: if your strategy is expressed as growing a percentage, you’re pursuing diminishing returns. If it’s expressed as strengthening your product value prop, you’re pursuing increasing returns.
That is the piece XO was missing — and the reason a feature-and-bug backlog quietly fails a SaaS company. The backlog answers the customers in the room. It moves percentages: of the customers who do this, how many can we also get to do that. It is pointed at exactly the customers you’re already over-indexed on. Nothing in it is built to ask the question that compounds: does this make the product more attractive to buyers? In this world, percentages still matter. They just stop being the point. Step-change versus competitors is the point.
XO measured the job. But in SaaS the job isn’t what compounds — the win is. So I needed the framework to track strengths the way increasing returns work: a strength where the better you get at it, relative to every other offer on the table, the more deals it pulls in. Not a feature that’s merely good. A strength with no ceiling on how many buyers it can win.
That is what a Winning Product Experience is. A WPX is an increasing-returns hypothesis: a bet that this is a strength where pulling further ahead of competitors keeps pulling in more deals. An EPX is its shadow — the experience that loses the deals a WPX should have won.
That split is WPX and EPX.
WPX and EPX: both sides of the coin
A Winning Product Experience (WPX) is what makes a buyer choose you and stay. Not what you sell — what they experience. It is a promise the product keeps. It is pull: the reason the next purchase, and the next, happens at all. It is where and how you win. There is no natural upper limit to how attractive this attribute can become.
An Eroding Product Experience (EPX) is what a customer experiences, or a buyer worries about, that wears trust down — what loses deals, blocks expansion, and drives churn. It is the place that same promise breaks. It is the leak in the pull.
Three things about that pairing carry the whole framework.
First, the bar for whether it’s a real WPX: your product has to be 10x better than the competition on this attribute. If your product isn’t 10x better on your WPX, you’re typically not winning on your product — you’re winning on something else, usually distribution. That’s a legitimate way to win, but it leaves you exposed and it caps your growth.
“10x” probably sounds too subjective to be useful, but it isn’t — you just have to know where to look for it. The 10x is in the attribute: judged head-to-head, your WPX is in a different class than the alternative. But that is not what you’ll see in the win data. There, a real WPX shows up as roughly a two-to-one preference — because the buyer is never choosing on one attribute alone. They are weighing your whole offer against the competitor’s, and the competitor’s own strengths pull the number back toward even. Two-to-one isn’t a weak 10x; it is what a 10x attribute looks like once it’s competing inside a real product, against a real rival.
For example, in consumer beverages, Pepsi could win the blind taste test, and Coke still won the sales two to one — because Coke’s WPX was never flavor, it was the brand story, and the brand story was genuinely 10x. And the ratio is a live reading, not a trophy. When that preference compressed toward three to two, that was not noise — that was the 10x decaying, in part because Pepsi eroded it through aggressive brand storytelling of their own. The WPX was sliding toward table stakes. The number moving is your earliest warning that a WPX is starting to fade.
And the second half still has to hold: the segment that cares about that attribute has to be growing. A two-to-one preference in a shrinking segment is a WPX with an expiration date.
The second element is voice — inherited straight from XO’s discipline. A WPX is never “we have easy-to-configure views.” It is the customer saying, in the first person: “we collect updates the way each team wants, and it’s all the same data underneath.” Same capability, completely different sentence — and only the second one tells you whether the thing you shipped is being experienced as a win. XO insisted on whats over hows. WPX/EPX insists on customer voice over product features. You cannot write a good one from inside the building, and that is the point: the discipline of customer voice drags your attention off your own roadmap and back onto the market.
The third is the word eroding. I chose it deliberately over “losing.” Losing names an event — a deal, a logo, a moment. Eroding names a direction. Direction is the whole game in a business of increasing returns, because pull compounds in both directions: the same dynamic that makes a strong offer stronger makes a leaking one leak faster. Read SaaS cancellation interviews honestly and customers rarely cite a single catastrophic failure. They cite an accumulation — value that never quite materialized, adoption that stalled, friction they got tired of absorbing. Trust doesn’t snap. It wears. An EPX names the wear while it’s still happening, which is the only window in which you can still reverse it.
The anatomy of a strong entry
A framework’s power lives in the bar for what qualifies. A strong WPX or EPX has four properties, and you’ll hear the XO DNA in all of them.
It’s in customer voice. First person, the words a buyer would actually use, ideally close to verbatim. If it sounds like a release note, rewrite it.
It’s evidence-anchored. Every entry traces to a source — a review corpus, a research study, a closed-deal interview. The highest-confidence entries come from closed-deal interviews, because that’s the moment a customer had nothing to gain by being polite. An entry with no source is an opinion in a costume.
It’s scoped. Specific enough to be measurable, broad enough to be a pattern. “The export button is gray” is too small. “The product is hard to use” is too vague. The right altitude is a recurring, nameable experience.
It’s shippable against. Every WPX and EPX must point to something you can invest in or fix. Pure outcomes — “trust,” “market position,” “social proof” — are not entries. They are downstream effects of the pull, not the pull itself. This is the rule that does the most work, because it keeps the framework from collapsing into a sentiment dashboard. WPX and EPX are prioritization inputs with a customer’s voice on the front and an investable capability on the back.
Themes have both
A theme rarely produces just a WPX or just an EPX. It produces both — because the same dimension that wins customers is the one that, pushed too far or not far enough, also erodes them. The WPX and the EPX aren’t a matched pair you assemble; they are the two things one capability inevitably generates.
Configuration is the clearest case. The very thing one buyer loves — “I can shape it to exactly how we work” — is the thing the next buyer struggles with — “there are too many knobs,” or “it still won’t bend the way I need.” Same capability, same theme, opposite faces. Turn the dial up and you get the WPX. Turn it down and you get the EPX. One dimension, read in both directions.
Take an all-in-one connected workspace — the kind of product that folds docs, wikis, databases, and project tracking into one tool. Read its themes as dials:
Theme — One flexible canvas WPX: “Docs, our wiki, the project tracker — it’s all one tool now. We killed three subscriptions and nothing lives in a silo anymore.” EPX: “The blank page is the problem. Everyone builds it differently, so six months in nobody can find anything and the workspace is a mess.”
Theme — Databases as a building block WPX: “Our roadmap, CRM, and content calendar are all just database views of the same data. I stopped copy-pasting between tools.” EPX: “Relations and rollups have a real learning curve. Most of my team treats it like a doc editor and never touches the power underneath.”
Notice the tell of a real EPX: it names the promise before it names the break. An EPX that doesn’t acknowledge the corresponding strength is usually just a complaint. An EPX that does is roadmap input — it tells you exactly which dial your customers are asking you to turn, and which way.
When a theme shows both faces clearly, you are not looking at two unrelated tickets. You are looking at one capability that is simultaneously your reason-to-buy and your reason-to-leave — a single lever on the pull of the whole offer. A feature-and-bug backlog will never surface that, because it filed the win and the erosion in two different folders and pointed both at the customers you already have.
How this changes the roadmap
Here is where WPX and EPX stop being a vocabulary and start being an operating model. The change is in what the roadmap is made of.
A bold product roadmap doesn’t allocate based on feature requests. It allocates against the WPXs and EPXs the product leader has chosen to strengthen or mitigate — and feature requests are evaluated only by how well they accomplish one or the other. A request doesn’t earn a slot by being loud or frequent. It earns a slot by demonstrably advancing a WPX or mitigating an EPX. The leader still chooses; nothing here is automatic. But the choice is now made in the open, against the experiences that win and erode deals, with a deliberate and heavy bias toward investing in the WPX over patching the EPX.
That bias rests on a hard truth: you cannot fix every EPX, and you shouldn’t try. Every strength has a counter. Customization is the tax on simplicity; depth is the tax on approachability. Chase every EPX to zero and you’ll erase the very WPX that made you worth buying. So the goal was never zero erosion. The goal is to make sure your WPX grows faster than your EPX leaks — you mitigate erosion, you don’t eliminate it. A strength that’s leaking is still a strength, as long as it’s widening faster than it’s draining.
In practice it lives in a single tracker — the entries the leader has chosen for the current cycle, with the unchosen ones still visible so the choice, and the bias, stays honest. Continuing with the connected-workspace product:
| ID | Type | Theme | Experience (customer voice) | Source |
|---|---|---|---|---|
| WPX-1 | WPX | One flexible canvas | “It’s all one tool now. We killed three subscriptions and nothing lives in a silo.” | Win/Loss |
| EPX-1 | EPX | One flexible canvas | “The blank page is the problem. Six months in, nobody can find anything.” | Reviews, Research |
| WPX-2 | WPX | Databases as a building block | “Roadmap, CRM, content calendar — all views of the same data. I stopped copy-pasting.” | Win/Loss |
| WPX-3 | WPX | Real-time collaboration | “Discussion happens on the doc itself. We cut a chunk of our meetings and Slack threads.” | Win/Loss, Reviews |
| EPX-2 | EPX | Offline & reliability | “If my connection drops, I can’t work. For a tool I live in all day, that worries me.” | Research |
Three things this tracker does that a backlog can’t. The two strongest elements share a shape — the flexibility and the database primitive are exactly what wins the deal and exactly what overwhelms the user, the same capability seen from both ends; the cycle decision reads straight off that, invest to widen the win and mitigate so the erosion never drowns it.
The same lens points outward. Your competitors have their own WPXs and EPXs, and you read theirs the way you read your own. Where a competitor has a WPX, you ask whether you can turn it into table stakes — meet it well enough for most buyers that it stops being a reason to choose them, and leave them serving only the extreme edge of the market. Where a competitor has an EPX, you make sure the buyer sees it clearly. Their leak becomes part of your pull.
When your WPX stops winning
You know it’s time to change when your WPX stops winning customers — when the reasons your target buyers say yes no longer feature your WPX front and center. When that happens, one of two things is true: you have to shift the market, or your gap was never strong enough to begin with. Working out which comes down to three reasons the market hasn’t been trained to value your WPX.
The first: you haven’t made it real. You haven’t leaned all the way in — in marketing, in positioning, in the sales narrative — to show why your WPX is a step-function better than the alternatives. The capability might genuinely be 10x, but if buyers can’t see it, it can’t pull them. The fix isn’t a new WPX. It’s committing fully to the one you have.
The second: your EPX is screaming. Your WPX is doing its job — buyers are interested — but deals fail earlier in the cycle because something else in the value proposition is actively pushing them away. An EPX gap drowns out the pull. The fix is to mitigate the EPX that’s losing the deal, not to chase a new WPX. Adding more pull does nothing when a weakness is screaming this loud.
The third: a competitor has a WPX that outpaces yours. The response is twofold. Press their most vulnerable EPX — can they handle enterprise scale, do they offer real configuration — and make buyers see it. But the more important move is to turn their WPX into a product attribute, a box that gets checked, that satisfies most buyers and leaves the competitor pursuing only extreme users.
Always carry a conjecture about the next WPX
The third reason is the one you can see coming, and it’s why you should never be without a conjecture about how you intend to win in the future.
Every WPX decays. What wins deals today becomes table stakes tomorrow. Microsoft Teams won first because it was bundled, then because it was a plausible web-conferencing option in the time of COVID, and finally because of its collaboration-centric way of working. The WPX shifted underneath it more than once — sometimes pulled by market opportunity, sometimes pushed by product leadership, sometimes settling into a durable product or brand attribute. None of those shifts announced themselves with much warning, and discovering the next WPX takes far longer than the warning you’ll get.
So the WPX list is not only a description of where you win today. It is also forward-looking on purpose. You deliberately seed it with WPXs you do not win now — experiences where you hold a theory, or a strategic intent, to win tomorrow. That is how you give future buyers a seat at a table where they otherwise have no voice. The work isn’t to wait until your current WPX fades and then scramble. It’s to always hold a live, testable bet on the next attribute you can be 10x better at — so the shift becomes a move you choose, not one a competitor forces on you.
Your brand is the most durable WPX
One last point, and it’s the most durable one: your brand is the best WPX you can hold.
Apple has held a design WPX since its inception, and because it has never wavered, that WPX stays durable and stays 10x ahead. Samsung might ship a beautifully designed product — but the moment it launches something that misses that bar, or doesn’t integrate seamlessly, it undermines the very claim it would need to make. A WPX holds only as long as you never compromise it. Brand is what a WPX becomes when you’ve refused to compromise it for long enough that buyers stop checking.
Strip away the acronyms and this Viewpoint is one argument. Steven Herbst’s XO taught a generation of us to stop shipping features and start delivering measurable outcomes. It was the right correction, and I’d build on it again tomorrow. But XO measures whether you deliver the outcome — and in SaaS, delivering the outcome is not the same as winning. A company gets quietly captured by its current customers, its backlog fills with their requests, and it optimizes — diminishingly, faithfully — for the people already in the room while the market moves under everyone.
A SaaS product has to be purchased for use and purchased for purchase, and it has to keep winning both, because the buyer’s gate never closes. WPX and EPX exist to keep the product pointed at that fight. A WPX names, in the customer’s voice, where the offer is pulling. An EPX names where the pull is leaking. Build the roadmap from those, bias hard toward the wins, mitigate the erosion enough that it never drowns them, and always carry a bet on the win you don’t own yet.
That is a bolder way to run a product organization. And in a business where the customer can leave at the next renewal, it is also the only one that wins.
If you run a version of WPX/EPX with your team, I’d like to hear how you adapted it — especially which paired themes surfaced first, and which dial they pointed to. Tell me what you find.