Every CRO has had this conversation. It is end of quarter. The forecast says you are going to land. Three deals everyone has been calling commit for six weeks suddenly slip. Nobody is shocked. Everyone, including the reps, knew "something felt off." The CRM doesn't know. The CRM still has all three deals in stage 4.

The standard diagnosis is "rep judgment failure" or "deal review discipline." Both are real. Neither is the root cause.

The root cause is that you are forecasting from data that lags reality by 4-12 weeks. By the time a deal moves from stage 4 to stage 5 in your CRM, the buying behavior that determines that move has already happened weeks earlier. You are measuring the shadow, not the object.

There is a forward signal. It has existed all along. It lives in the email threads, calendar invites, call transcripts, and LinkedIn activity that surround the deal. We call it relationship health. It is the most underused forecasting input in B2B revenue today.

What is relationship health?

Relationship health is the measurable trend in how a buying account is engaging with your team across all communication surfaces, scored against the historical pattern of accounts that closed.

Concretely, it has four components:

1. Response cadence. Are stakeholders replying faster, slower, or not at all relative to two weeks ago?
2. Two-way exchange. Is the conversation balanced, or has it become your reps sending and the buyer occasionally acknowledging?
3. Stakeholder coverage. How many distinct humans on the buying side are engaged this week vs. last week, and how does that compare to deals at the same stage that closed?
4. Champion mobility and activity. Is your champion still posting on LinkedIn about your category? Are they showing up to meetings? Did they just change jobs?

Each component is independently noisy. Together, they form a signal that leads CRM stage data by roughly 6-8 weeks in our customer cohort. That is the difference between forecasting a miss in week 4 of the quarter and being surprised by it in week 12.

Why CRM data is a lagging indicator

CRM stages move when reps update them. Reps update them when something verifiable has happened, usually a meeting, a document sent, or a verbal commitment.

Each of those events is itself the output of a buying behavior, not the behavior itself. A buyer agreed to a meeting because their interest moved. The interest moved because, somewhere upstream, a relationship moved. By the time the calendar invite lands, the relationship signal is 2-4 weeks old. By the time the rep updates the stage, it is 4-8 weeks old. By the time pipeline review picks it up, it is 8-12 weeks old.

This is the structural reason "deal slippage" feels like a surprise even though it shouldn't. The buying signal was there. The instrumentation wasn't.

A second problem: CRM data is incomplete by design. Roughly one third of the people in a typical buying committee never make it into the CRM at all. They are forwarded the email, sit in the meeting silently, or are introduced informally. They are real economic buyers. They are invisible to your forecast.

If you are forecasting from a system that is both lagging and incomplete, the question is not whether you will miss. The question is when.

The four forward signals, in order of strength

We've spent the last two years watching relationship signals across hundreds of deals at our customer cohort. A rough ranking, by predictive strength:

Signal 1: Response cadence delta

The single highest-leverage signal is the change in how quickly stakeholders respond, measured as a delta from their personal baseline.

A VP of Engineering who replies in 4 hours on Tuesday and 36 hours by Thursday is not "busy." They are disengaging. By the time the lag hits a week, the deal is sliding.

This signal is invisible in CRM activity reporting (which counts emails sent, not response time) and noisy in raw email logs (because some people are just slow). Measured against the individual's own baseline, it is one of the cleanest predictive inputs you have.

Signal 2: Stakeholder coverage gap

The shape of your buying committee matters more than its size.

Compare two deals, both in stage 4, both $500K ACV:

• Deal A: 5 contacts engaged, including 1 economic buyer, 2 users, 1 IT/security, 1 finance.
• Deal B: 5 contacts engaged, all 5 from the same engineering team.

Deal A is statistically about 3x more likely to close than Deal B. Deal B looks identical in your CRM activity reports.

The gap that matters is between who is engaging and who needs to engage for this deal to close. Most buying processes require explicit alignment from 6-11 people. If you only have 5 of them and they're all in one function, you don't have a deal. You have a champion.

Signal 3: Two-way exchange ratio

A deal where your team sends 12 messages and the buyer sends 2 is in a different state than a deal where the ratio is 7-to-7, even if the total count is identical.

Two-way ratio is the closest thing to a measure of buyer-side initiative. When a buyer is asking questions, forwarding internal context, looping in colleagues, the deal is alive. When the rep is doing all the talking, the deal is on life support.

This is also the cleanest predictor of "no decision" losses, which are now the majority of lost deals in most B2B segments. No-decision losses don't fail because you lost to a competitor. They fail because the buyer's own commitment quietly drained.

Signal 4: Champion mobility and external activity

The leakiest part of most pipelines is champion attrition.

If your champion is interviewing elsewhere, your deal is in trouble three weeks before anyone tells you. If your champion just got promoted, your deal often accelerates (and an expansion conversation just appeared). If your champion left and joined a new company last week, your CRM has stale contact data on a deal that is about to die and a new opportunity at the new account that nobody on your team has been alerted to.

Champion mobility is a signal almost no GTM team is currently instrumented to act on. It is also one of the most expensive blind spots. We've seen a single customer (Narvar) generate $17 million in pipeline in one year just from tracking champion job changes and triggering a warm intro the day they land at the new account.

What you should do, if you're a CRO or RevOps leader

Three concrete moves, in priority order.

One. Audit your current forecast inputs. Pull the last four quarters of slipped deals. For each, look back 8-12 weeks before the slip and identify whether response cadence, stakeholder coverage, two-way ratio, or champion activity were already deteriorating. Most teams find that 70% or more of slipped deals had a visible relationship-health signal 8+ weeks earlier. That number is the size of your current blind spot.

Two. Add one relationship-health metric to deal review. Not all four at once. One. We recommend starting with stakeholder coverage gap because it is the easiest to measure and the easiest to act on (the rep can multi-thread). Report it weekly per deal, color-coded against the historical pattern of closed deals at the same stage.

Three. Connect champion mobility to your motion. Every customer of yours has a 25-30% annual churn-by-job-change rate. That means a quarter of your installed base is going to be in a new company in the next year. Each of those is a warm pipeline opportunity. Almost nobody operationalizes this. Instrument it now, even crudely (a Slack channel that fires when a contact's title changes is a start), and you will discover pipeline you didn't know existed.

Why this is the next category of revenue analytics

CRMs were built when revenue was a forms-and-fields problem. You wrote down what happened. You looked at the report.

The current generation of "revenue intelligence" tools (Gong, Clari, and others) added activity data on top: calls, emails, meeting participation. That was a big step. It is also not enough, because activity is still a measure of what you did, not what the buyer is doing back.

The next layer is the relationship signal itself: the trend in how the other side of the table is engaging, scored against the patterns that predict outcomes. This is the layer where the forecast becomes leading instead of lagging.

The companies that figure this out first will not "have a better dashboard." They will catch deals slipping six weeks earlier, multi-thread accounts that look healthy but aren't, and convert champion job changes into pipeline before competitors notice the contact moved. None of that is visible in stage data. All of it is visible in the relationship signal.

The CRM was the last fifteen years of GTM analytics. The relationship graph is the next fifteen.

Curious what relationship-health instrumentation looks like in your pipeline today? Run a free pipeline leak assessment and see which deals have signals you are not acting on.