SciFin's Survey Makes a Better Case for a Forecast Audit Than for Another Tool
SciFin, a startup selling revenue-context software, commissioned Censuswide to poll 500 U.S. sales and RevOps leaders. Just 27.2% said a lost deal's warning signs existed but went unconnected.
TL;DR
- SciFin's Censuswide survey of 500 U.S. sales and RevOps leaders found 78% surprised by deals their systems called healthy.
- The survey never links rising rev-tech spend to those surprises, and only 27.2% said a lost deal's warnings existed somewhere.
- The slip rate and the lag before a buyer event reaches the CRM, both left to leaders' impressions, already sit in the CRM's field history.
Seventy-eight percent of U.S. sales and revenue operations leaders say they've been surprised by a deal falling through that their systems showed as fine, according to The Cost of the Revenue Context Gap. In the same survey, 65% say their company raised annual spending on revenue systems this year.
SciFin commissioned the survey from Censuswide, and it sells software aimed at the problem the survey measures. Censuswide polled 500 leaders at companies with 500 or more employees and at least one quota-carrying seller between August 28 and September 2, per the methodology note in SciFin's Business Wire release of October 13. The full report sits behind a signup form.
SciFin is young. Nutanix co-founder Mohit Aron started it in 2024, and it launched publicly on September 1 with a $44 million seed round co-led by Altimeter and Madrona, so the fieldwork ran the same week as the launch. The release that published the findings also introduced Context Convergence, a product that merges structured and unstructured customer signals into one layer its agents and forecasting apps read from. Every number here is self-reported by leaders answering a vendor's questionnaire.
SciFin's survey never connects the spending to the surprises
The headline pairs two numbers as if one undercuts the other. Budgets went up, and the surprises kept coming. The survey doesn't test that link.
The 78% question asks whether a leader has ever been caught out by a deal that looked healthy, which describes almost anyone who has carried a forecast for a few years. The release reports no split of surprise rates by whether spending rose. Without that cut, the bigger budgets could just as easily belong to the teams that got blindsided most.
The firmer findings sit lower in the release. Leaders estimate that 35% of forecasted deals slip into a later quarter. Eighty-one percent say most CRM updates reflect what someone had time to type rather than what happened, and 77.2% say they regularly leave their forecasting tools to learn what's going on. Six in ten (61%) gave their board a number more conservative than the one they believed.
Outside data points the same way. Gartner's 2020 State of Sales Operations survey found only 45% of sales leaders and sellers had high confidence in their organization's forecast accuracy, and only 47% rated their company's data as high quality. Gong, which sells forecasting software, reported in 2022 that just 24% of sales leaders were confident in their team's forecast, from a survey of 928 revenue professionals.
In SciFin's version, 75% couldn't honestly tell their CEO or board they were "very confident" in this quarter's number. The questions differ, so the figures don’t make a trend line. Still, we couldn’t find a survey from the six years of revenue-intelligence buying since that shows confidence going up.
SciFin's own numbers also cut against its pitch. Just 27.2% said that when a deal was lost, the warning signs had existed somewhere in the organization and nobody connected them in time. A context layer is built to catch that case, and fewer than three in ten respondents described it.
Only 7.2% blamed dishonest reps for rosy forecasts, while 25.2% pointed to natural optimism and 21.6% said flagging risk earlier hadn't changed outcomes in the past. If the warning often never existed, or raising it changed nothing, more signal plumbing won't fix the forecast. Commissioned research tends to put its weight on value that’s hard for a buyer to verify, which is what Outreach's Forrester study did with modeled sales gains.
A context layer needs a warning to connect. Most leaders didn't report one.
The survey's limits apply to every vendor selling into the 78%, Gong and SciFin among them. They also apply to a RevOps budget already shifting toward governing the data model rather than buying point tools.
The survey left two numbers to leaders' impressions, the slip rate and the lag between a buyer event and the CRM field that records it. Both sit in the CRM's own field history. Against the 78% surprised by a deal their systems called healthy, just 27.2% said the warning signs had existed somewhere in the organization and nobody connected them in time.
If the warning often never existed, or raising it changed nothing, more signal plumbing won't fix the forecast.
The figure that would settle it is the slip rate in CRM field history, which leaders in SciFin's survey estimated at 35% of forecasted deals.
