Sales Cycles Now Outrun Your Attribution Window

B2B SaaS sales cycles have gotten materially longer, and most attribution setups have not kept up. The median cycle now runs about 84 days — up roughly 22% since 2022 — with a mean nearer 134 days once enterprise deals drag the average out, according to 2026 SaaS sales-cycle benchmarks. Yet the lookback windows most teams actually run — inherited platform defaults of 7, 30, or 60 days — were sized for a shorter era. When the cycle outgrows the window, the window silently deletes the touches that started the deal, and your channel report quietly stops describing your funnel.

This is a different problem from picking an attribution model or chasing the dark funnel, and it deserves its own treatment because the root cause is a moving number: your sales cycle is getting longer while your window stayed fixed. This guide covers how long cycles actually are in 2026 and why they grew, the mechanical way a too-short window erases the top of your funnel, how to tell this failure apart from the dark funnel, and how to size a window from your own cycle data so the report stops lying about which channels work.

How long the cycle actually is now

The headline number is an 84-day median, but the segment breakdown is where the measurement problem lives. SMB deals under $10K ACV close in a median of around 27 days, mid-market ($10K-$50K ACV) runs 45-75 days, and enterprise (>$50K ACV) runs 90-180 days — with $250K+ deals often stretching to a year. A single blended window cannot fit that spread: whatever number you pick is simultaneously too long for the SMB tail and too short for enterprise. The mean of 134 days versus the 84-day median is itself the tell — the gap between the two is the enterprise tail, and that tail is exactly the set of deals a short window truncates.

The lengthening is structural, not a blip, which is why it will keep drifting past static windows. Buying committees have grown to six to ten stakeholders, CFOs now gate most deals above $50K, and security due diligence adds weeks that never used to exist. None of that reverses next quarter. The consequence for analytics is that any window you set from memory or from a 2022-era rule of thumb is already too short, and the error grows every quarter the cycle keeps stretching. Measurement that assumes a stable cycle length is measuring a business that no longer exists.

Why a hard cutoff erases the top of your funnel

The damage comes from a specific property of the lookback window: it is a hard cutoff, not a gentle decay. As covered in the lookback-window sizing guide, a touch older than the window boundary receives zero credit, not reduced credit — the system deletes it rather than down-weighting it. So when an 84-day deal closes and your window is 30 days, every touch from the first 54 days of that journey is invisible. The blog post that introduced the product, the comparison search a month out, the webinar that built trust — all gone, and the entire conversion is credited to the demo request or branded search at the very end.

The bias is directional, which makes it dangerous rather than merely noisy. It does not spread error randomly; it systematically transfers credit from top-of-funnel channels to bottom-of-funnel channels. Content, organic social, and awareness campaigns — the work that seeds the deal — get zeroed out, while retargeting and branded search look like your best performers. Act on that report and you defund the pipeline-filling channels, then watch the pipeline dry up a quarter later once the seeding stops. This is the same class of invisible measurement failure described in the acquisition signal-decay guide: nothing breaks loudly, the numbers just quietly stop being true.

This is not the dark funnel — do not confuse them

It is easy to file this under "attribution is broken" alongside the dark funnel, but they are distinct failures with distinct fixes, and conflating them leads to solving the wrong one. The dark funnel is a collection gap: word-of-mouth, private Slack communities, and peer recommendations never leave a trackable touch, so a large share of pipeline shows up as "direct" or unattributed. The sales-cycle-versus-window mismatch is the opposite — the touch was captured correctly, it just aged out of the window before the deal closed. You are throwing away data you already have.

The distinction matters because the remedies do not overlap. Widening your lookback window recovers real, tracked touches that were being discarded; it does nothing for the dark funnel, which needs self-reported attribution or "how did you hear about us" survey data to surface channels that leave no digital trace. A team can suffer both at once — a too-short window hiding the content that was tracked, and a dark funnel hiding the community that was not — and end up misreading its channel mix twice over. Diagnose which one you have before reaching for a fix, or you will buy a survey tool to solve a window setting, or vice versa.

Measure your real cycle, then size the window to it

The correct window is a measurement, not a default. Pull your closed-won deals from the last 6-12 months and, for each one, compute the days between the first recorded marketing or product touch and the close date. Use first touch, not CRM opportunity-created date — the opportunity stage starts long after the journey began, so measuring from it reproduces the exact truncation you are trying to fix. Then plot the full distribution and read the 50th, 80th, and 90th percentiles. The median tells you the typical deal; the 90th percentile tells you how far back the window has to reach to cover the slow deals you most want to keep.

Set the window to the 80th or 90th percentile of that distribution, and do it per segment. For an 84-day median with a heavy enterprise tail, that commonly lands well past 90 days — often 120-180. Segmenting matters because one blended window over-credits stale touches on fast SMB deals while still chopping the enterprise tail; a short window for SMB and a long one for enterprise fits both. This is the same discipline of defining a metric from data rather than habit that keeps conversion-window definitions honest, applied to the touch-crediting side of the funnel.

What to fix first, and what it connects to

Start with an audit, because most teams do not know what window they are running. Check each ad platform, your product analytics tool, and any warehouse attribution model separately — they frequently carry different windows, which is one common reason the same deal is credited to different channels in different dashboards. Simply documenting the current window per system, next to your freshly measured cycle-length percentiles, often explains a reporting discrepancy that has been blamed on "bad data" for months. If your longest window is shorter than your 80th-percentile cycle, you have found a concrete, fixable source of misattribution.

Then standardize and revisit. Apply a cycle-driven window consistently where you can, record it alongside the attribution model in your metric definitions, and recompute the cycle distribution quarterly — if the 2026 trend holds, the cycle keeps lengthening and last year's window keeps getting shorter in relative terms. The window is only one of two independent knobs, so pair this with the attribution model selection guide to decide how credit is split among the touches the wider window now lets through. For long enterprise motions, the post-close stages matter too — the bowtie-funnel instrumentation guide extends the same time-aware thinking past the first purchase into expansion.