B2B sales cycle length by industry, and why yours differs

Somebody always wants a benchmark to compare against. The published ones are worth reading and worth distrusting in equal measure, and the gap between them and your own number is the useful part.

The question usually arrives after a bad quarter. Deals took longer than anyone planned for, the forecast slipped twice, and somebody asks what a normal sales cycle is supposed to look like in our industry. It is a fair question and there is an answer, but the answer is less useful than the reason your number is different from it.

This is what the published benchmarks say, what they quietly leave out, and how to work out the figure you should actually be planning against.

What the published benchmarks say

The most complete public table at the moment comes from a Focus Digital study published in June 2026, which puts average cycle length by industry as follows.

Industry	Average days
Retail	70
Hospitality	85
Software	90
Financial services	98
Consulting	103
Automotive	104
Real estate	105
Logistics	117
Technology	121
Healthcare	125
Education	126
Insurance	127
Manufacturing	130
Construction	134
Pharmaceuticals	153
Energy	155
Non-profit	162

A separate dataset is more useful because it says how it was built. The Optifai pipeline study covers 939 B2B software companies with stage level CRM data between the second quarter of 2025 and the first quarter of 2026. It reports a median of 84 days across all segments, and it reports ranges rather than single figures because it publishes the 25th to 75th percentile of won deals.

Its split by deal size does more work than the industry table does:

Segment	Range	What is different about it
Under 15k annual contract value	14 to 30 days	One decision maker, often no procurement at all
15k to 100k	30 to 90 days	Procurement appears, two or three stakeholders
Over 100k	90 to 180 days and up	Committee, security review, multi quarter budget

Read the methodology before you use the number

The industry table is the one people screenshot, and it is the one to treat most carefully. The published report names its own research as the source and lists other 2026 benchmarks alongside it, but it does not disclose a sample size, a collection method or any statistical validation. That does not make the figures wrong. It means you cannot tell how wrong they might be, which is a different problem and a worse one for planning.

The Optifai figures are more defensible because the sample is stated, but they carry two limits that matter. The sample is business software companies, so a logistics or construction seller is reading a number produced by a different kind of business. And the percentiles are calculated on won deals only, which is the single most common distortion in this whole category of statistic.

Won-deals-only is worth sitting with. Your slowest opportunities are disproportionately the ones that never close. If you measure only the deals that landed, you are measuring a sample that has already had its worst cases removed, and every benchmark built that way reads faster than the reality a rep experiences.

What actually makes a cycle long

Industry is a weak predictor. It correlates with cycle length mostly because it correlates with the things that genuinely cause it, and those are worth naming directly.

• Deal size, which sets everything else. Price is what summons procurement, legal and a second approver. The Optifai split above is really a chart of how many people have to say yes.
• The size of the buying group. The same study puts the average at 6.8 stakeholders, against 5.4 previously. Every added stakeholder adds a schedule to coordinate, not just an opinion to win.
• Whether a security or compliance review is triggered. This is close to binary. It either happens or it does not, and when it happens it adds weeks that no amount of selling technique removes.
• Whether the budget already exists. A deal that needs new budget is not slow because the buyer is slow. It is waiting for a planning cycle, and no follow-up sequence shortens a planning cycle.
• Whether the buyer has a deadline. A renewal, a regulation, a launch. Deals without one drift, and that is the honest reason behind most of the stalling we describe in why B2B deals stall.

How to measure your own, properly

Most teams that quote a cycle length are quoting a number nobody can reproduce, because the start date was never agreed. Fix that first and the rest is arithmetic.

Pick a start event that exists in the record and does not depend on judgement. First meeting held is the best available candidate for most teams. First touch is worse, because it includes however long the prospect sat in a sequence, and that is an outbound cadence measurement rather than a sales cycle. Opportunity created is worse still, because reps create opportunities at whatever moment their manager rewards.

Then measure to close, both won and lost, and report the median rather than the mean. One nine month enterprise deal will pull an average far enough to make it useless for planning, and the median will not move. Report the two segments separately if you sell to both small and large accounts, because a blended figure describes no deal you have ever run.

Doing this weekly by hand is where definitions quietly drift, since whoever exports the data makes a small judgement call every time. Keeping stage timestamps trustworthy is mostly a discipline problem rather than a tooling one, covered in CRM hygiene for small sales teams. If it is genuinely a tooling problem, Rivl set out when a bought system beats a built one in their build versus buy read on custom CRM cost.

What you can shorten and what you cannot

The parts of a cycle that respond to effort are the parts you control: how fast you get the second meeting, how quickly the right people are in the room, whether the proposal answers procurement's questions before procurement asks them. Compressing those is real work and it produces real weeks.

The parts that do not respond are the buyer's budget calendar, their security queue and their legal backlog. Pushing on those does not shorten anything. It converts a slow deal into a lost one, and it is worth being honest with a team about which category a specific delay falls into before setting a target on it.

The lever nobody counts as a lever is qualification. A cycle length improves substantially when you stop opening deals that were never going to close this year, and the mechanism there is set out in a B2B lead qualification framework that survives an audit. Knowing who is genuinely in the room is the other half of it, which is the subject of the B2B buying committee.

Honest limits

If you close fewer than about thirty deals a year, your own median is not stable enough to plan against either. Two unusual deals will move it. At that volume the better instrument is reading each open deal individually, which a small team can genuinely do and a large one cannot.

And a cycle length is a description, not a target. Teams that are set a target on it tend to hit it by closing the fast deals and quietly abandoning the slow ones, which shortens the reported number while shrinking the business. If you are going to measure it, measure it alongside win rate and average deal size, or you have built an incentive to sell less.