Pipeline is behind plan. Leadership wants a response. The suggestions arrive almost immediately: spend more, add a channel, rewrite the campaigns, launch outbound, replace the agency, buy a platform.
Some of those actions may eventually be right. But over 15 years of running demand generation in B2B SaaS, including nearly nine years inside a company scaling from roughly $50M to $450M ARR, I’ve learned that the most expensive mistakes usually start the same way: a correct sense of urgency and an incomplete diagnosis.
So before I change anything in a demand program, I work through six questions. Not as a workshop exercise. As the fastest route to knowing which change is actually worth making, and in what order.
By the end, I want a defensible view of the commercial target, the priority market, the reliability of the evidence, the largest conversion constraint, what sales needs to be able to execute, and who owns the corrective work. Here’s how I get there.
1. What commercial outcome is this program expected to support?
The first question sounds obvious and rarely has a clean answer. What revenue, pipeline, or opportunity number is marketing actually on the hook for? Sourced or influenced? Over what period? New logos or expansion? And is the target even achievable given the market, budget, sales capacity, and time available?
I’m not looking for a perfect financial model. I’m looking for the assumptions connecting the business goal to the amount and type of demand required, because a channel plan cannot compensate for an undefined or internally inconsistent mandate. If the target assumes a win rate, sales cycle, or ACV that nobody can defend, that’s the first finding, and it changes everything downstream: how much pipeline is needed, which segments get the budget, and what “working” even means.
I’ve had to do this modeling honestly enough to tell a CEO that a target wasn’t reachable from the current motion alone, and to show what the realistic floor was and what would have to be added to close the gap. That conversation is uncomfortable exactly once. Building a program against a number nobody believes is uncomfortable forever.
2. Which customers and segments matter most?
Most companies have an ICP document. Far fewer have a usable market decision, which is a different thing. The test: does the segment definition actually change targeting, budget, campaign design, routing, and seller behavior? If every segment gets roughly equal attention, there is no priority, just a list.
The questions I work through: which segments produce the strongest commercial outcomes, not just the most leads? Where does the company have credible differentiation? Which segments can sales actually support? And, hardest of all, what will the company stop treating as an equal priority?
The evidence that matters here is conversion data by segment, actual customer economics, and what sellers say about the deals they win and lose. When those three disagree with the stated ICP, believe them over the document.
I’ve watched this go wrong from the org chart down. At one scale-up, roughly 90% of the customers sat in the smaller-business segment while roughly 90% of the revenue came from enterprise. Leadership read that mix as a paid-allocation problem: too much spend acquiring small businesses, not enough winning large ones.
The assumed fix was to split budgets and sales effort between two new business units, divide each channel in roughly the same proportions, and separate campaigns so rigidly that audiences became too small, budgets were spread across duplicative programs, and the platforms had less data from which to optimize.
But the two segments’ buyers behaved nothing alike. Smaller businesses had much shorter sales cycles and responded well to search and other direct-response programs. Enterprise buyers moved through far longer, more complex journeys, bought through committees, leaned on analyst research, and needed ABM, longer nurture, and coordinated programs across channels.
The immediate result was higher CPA on Google and LinkedIn and a sharp decline in smaller-business and mid-market lead volume, even though that acquisition motion had been working.
Lead volume recovered when we simplified the Google and LinkedIn campaign structures, removed duplicative campaigns that existed mainly because of the organizational model, and restored enough audience and budget scale for the platforms to operate efficiently. We also improved ad messaging, landing pages, and website content around the needs of the actual ICPs.
The budget model changed with it. Instead of assigning each business unit an equal share of every channel, we allocated investment according to how different buyers researched, evaluated, and purchased. Search and other direct-response channels continued to play an important role for smaller businesses, while enterprise investment shifted toward ABM, longer nurture programs, analyst influence, and coordinated campaigns across channels.
The commercial objective had been valid. The operating model used to pursue it had treated very different buying journeys as though they required the same channel mix.
3. Can the funnel data be trusted enough to make the decision?
Before I compare channels or reallocate budget, I need to know whether the numbers can carry the weight of the decision. Are lifecycle and opportunity stages consistently defined? Are source and campaign fields complete enough to compare anything? Do marketing, sales, and finance mean the same thing when they say “qualified”? Can sourced and influenced pipeline be separated?
The standard is decision-grade evidence, not artificial precision. Perfect data doesn’t exist. What a senior leader needs is the ability to say clearly: this is what the data supports, this is what it suggests, this is what it cannot prove, and these are the decisions that have to wait for better instrumentation.
What decision-grade looks like in practice is reporting built for the altitude of the decision. At scale, I ran this as three tiers: Salesforce dashboards for the sales organization covering pipeline performance, lead sources, and volume; Power BI reporting for executive leadership tracked against the business objectives themselves, marketing-originated and marketing-influenced pipeline against revenue targets; and channel-level dashboards for marketing leadership, updating on daily through annualized cadences from web, pipeline, and advertising data. Different audiences, different depth, one set of definitions underneath. When the definitions hold across all three tiers, the numbers can settle arguments. When they don’t, the dashboards just automate the disagreement.
I’ve spent years working with influenced-pipeline measurement at meaningful scale, and the lesson that survived all of it is simple: large pipeline numbers are only useful when the definitions, lookback windows, and use cases stay explicit. The moment “influenced” quietly becomes “attributed” in a leadership deck, the number stops informing decisions and starts defending them.
4. Where does conversion weaken, and what are the economics?
This is where most tactical debates should actually live, and mostly don’t. Channel performance is partly produced by the system around the channel. A “failing” paid program might be a targeting problem, a landing page problem, a qualification problem, or a sales follow-up problem wearing a media-metrics costume.
So I look for the largest meaningful drop: audience, response, website conversion, qualification, meeting creation, opportunity creation, progression. I compare segments and channels using equivalent definitions. And I ask what the company actually pays for a qualified action, an opportunity, a pipeline dollar, because that’s the language budget decisions get made in.
The clearest example from my own work: early in my time owning web and digital at a compliance SaaS company, paid performance looked like a channel problem. The diagnosis said otherwise. The constraint was the conversion system around the channels: pages that didn’t continue the promise the ads made, paths that leaked intent, campaigns misaligned to the pages receiving them. We fixed the system first. Within the first six to nine months, website conversion doubled, time on site tripled, and average CPA fell from roughly $1,200 to $400 in Google Ads and from roughly $2,200 to $800 in LinkedIn Ads. The improvement came from fixing the conversion system surrounding the channels rather than treating media performance in isolation. The numbers behind that work are on my results page.
Segment expansion ran into the same system truth. The company’s content was built for financial services buyers, in US English, which made adjacent segments expensive to reach: you cannot point public-sector or EMEA campaigns at pages written for someone else and expect the channel to perform. Rather than rebuild the site for each segment, I built more than 90 website personalization experiences that dynamically replaced text, CTAs, links, and images based on who was visiting, using reverse-IP lookup mapped to first- and third-party data. Bounce rates for those two segments fell 47%. For those audiences, the website experience had become a meaningful constraint on campaign performance.
5. What happens after demand is created?
Marketing performance conclusions are only as valid as the execution downstream of them. Who receives the lead or meeting? What qualifies it for action? How fast does follow-up happen, and does the seller get enough context to act well? What happens when routing fails? Is there enough sales capacity to work what’s being created? And how does what sellers learn flow back into targeting and messaging?
I’ve built this layer from scratch in zero-to-one programs: lifecycle and status definitions, routing and ownership rules, follow-up expectations, seller guides, the reporting that shows whether any of it is happening. What that work teaches you is that a channel can look broken for months when the actual failure is sitting between the form fill and the first call. If you don’t inspect this layer, you will misdiagnose the ones above it.
This isn’t marketing managing sales. It’s establishing the commercial requirements of the handoff and agreeing on shared expectations, so that when a channel number moves, everyone knows whether marketing or execution moved it.
6. Who owns the decisions and the corrective work?
Most underperforming programs I’ve seen had enough analysis. What they lacked was decision ownership. Who owns the commercial target? Who can change routing and seller expectations? Who is responsible for fixing the known data problem everyone mentions in every meeting? What can the demand leader decide independently, and what needs the CRO, RevOps, or finance in the room?
A diagnosis that doesn’t end in named owners, dates, and an operating cadence is just a delay with better formatting. This question is also where I find out whether the role itself is set up to succeed: if the person accountable for pipeline can’t influence routing, definitions, or budget allocation, the program has an org problem before it has a marketing problem.
What the diagnosis should produce
This is the part that separates a diagnostic from a checklist. A completed first pass should produce a short, concrete set of artifacts:
The commercial mandate, with its assumptions written down. The priority segments, including what’s being deprioritized. An honest confidence assessment of the data: what’s reliable, what’s directional, what’s unusable. A constraint map showing the most consequential breaks across acquisition, conversion, sales execution, and measurement. The immediate actions that are reversible or high-confidence enough to start now. The foundation work that has to happen for durable improvement. Named owners and a review cadence. The decisions deliberately deferred until evidence improves. And a 30-60-90 sequence that turns all of it into an operating plan rather than a recommendation list.
That last artifact is the answer to the fair objection that diagnosis delays execution. It shouldn’t. Productive in-flight work keeps running, reversible improvements start immediately, and the diagnosis runs in parallel against the highest-risk assumptions. What it prevents is the expensive kind of speed: confidently fixing the wrong thing.
The questions don’t change. The depth does.
At a small B2B SaaS company, this diagnostic is fast and pragmatic: basic lifecycle definitions, a usable report, focused channels, disciplined follow-up, and visible progress while the foundations get built. At mid-market, it becomes a portfolio question: segment and regional economics, team and agency capability, common definitions, competing stakeholder priorities. At enterprise scale, it’s substantially about decision rights, coordination across business units, attribution governance, and the sequencing of change itself.
I’ve operated at both ends of that spectrum, and the six domains held up everywhere. What changed was how deep each one had to go and who needed to be in the room.
Diagnosis creates the basis for responsible action
I don’t expect every question to have a perfect answer before work begins. I do expect the assumptions, the evidence gaps, the priorities, the owners, and the next decisions to be visible. That’s what turns a request for more pipeline into something a revenue team can actually operate.
This diagnostic approach is part of how I evaluate and lead demand generation programs across stages of growth. You can see the results it has produced, or read more about how I approach demand generation.