MES forecasts: tie pilot and capex stages to the buying panel

How long is the sales cycle for MES software?

Published estimates put a first MES deal at roughly 6–12 months from first contact to signature. None of them shows a sample of MES deals behind its range. Clean, which sells prospecting software, estimates 6–12 months and says its figures are not a survey.^↗ A&C Growth gives 180–300 days, drawn from its own campaign and deal analyses and published research.^↗ By our estimate, software deals with manufacturers take 4–12 months (Illustrative).^↗ The only large measured figure is cross-industry. In 6sense’s 2025 survey of nearly 4,000 buyers, the average buying cycle was 10.1 months, for a median purchase of $200–300k.^↗

A 9-month planning assumption sits inside all 3 ranges. It still won’t tell you which open deal closes this quarter. An MES sits between the ERP and the machines, so operations, OT engineering, IT and security all review it. A pilot, a security review and a capital request stand between a technical win and an order. By our estimate, an MES pilot runs 2–4 weeks, while budget or capex approval takes 4–12 weeks or waits for the next budget year (Illustrative).^↗ A median hides that spread.

We read the 12 pages AI answers cite for this question, and 11 could still be opened.^↗ Of those, 4 are buyer’s guides, 2 are accounting texts and only 2 give a cycle length.^↗ None ties a finance seat, such as the CFO or the plant controller, to a forecast stage.^↗

What should the pilot, capex and first-site stages require?

Each stage should count only when someone in the buying panel has done something you can see in the CRM. A rep’s confidence doesn’t count. The table sets out the stages we would use for a first-site MES or automation software deal, and the seat each one needs. Our manufacturing industry page maps the whole panel, from the COO to the works council.

Finance comes in before the shortlist, while the business case is still taking shape. Across 655,000 opportunities, Ebsta and Pavilion found win rates 55% higher when decision makers were actively involved in the first 2 stages.^↗ That benchmark isn’t specific to finance or to manufacturing. At a manufacturer, though, a capital request usually passes the plant controller or the CFO on its way to approval. So we treat finance as a decision maker from the start.

The pilot needs a charter before it starts: the line, the success measures with today’s values, the end date, the scale-up criteria and the first-site price. The operations owner who will decide after the pilot signs it. Without a charter, a pilot runs until someone stops it, and your pilot stage fills with deals that never age out.

The plant controller should see the business case before the pilot ends. The pilot produces the numbers for the capital request. If finance first sees them afterwards, the case starts again while the deal waits. A passed pilot is still only a stated intent. In a study of more than 2.5M recorded sales conversations, 40–60% of deals were lost to customers who said they meant to buy but didn’t act.^↗

“Engaged” needs a definition the CRM can check. Our build method defines it per company, for example as a two-way interaction in the last 90 days: a call, a meeting or a written reply.^↗ If you qualify with MEDDPICC, the economic buyer here is the approver at the first-site amount. The decision process is the capex calendar, and the paper process is your time from approval to order.

Forecast the first site on its own. Book a paid pilot as a small separate order. Treat each further plant as a new opportunity, because each plant often decides on its own P&L. The first-site rollout then sets when revenue starts, and that follows the plant’s shutdown window more than the contract date.

How do you date the capex approval?

Date it from the buyer’s approval calendar: who signs at the deal’s amount, when they next meet and whether the project is in this year’s capex plan. Keep that apart from whether the manufacturer will buy at all. A deal can be likely to close and still unlikely to close this quarter.

A worked example shows how (Illustrative: the vendor, the manufacturer and the calendar are fictional). An MES pilot on a single line ends on 20 November, and the first-site order is €180k.^↗ At this manufacturer, requests above €100k go to a capital committee that meets in the first week of each quarter.^↗ Next year’s capex plan closed on 30 September, without this project.^↗

The rep forecasts December, because the pilot passed. The calendar disagrees. If the plant controller submits the request by mid-December, the earliest approval is the January committee. The vendor’s own baseline from approval to order is 30 days, so the order lands in February.^↗ That makes it a first-quarter deal at best case, and it reaches commit only once the committee approves it. If the committee funds only planned projects, it waits for next year’s plan.

Record these dates on the account or the deal: the fiscal year-end, the date the capex plan closes, each approver’s limit, the next approval meeting and the shutdown window. Then make every close-date move carry a reason. Our audit scores slipped close dates 1 out of 5 when more than 40% of open deals have had their close date moved twice or more.^↗ The same discipline holds for govtech deals that wait on a council vote.

How do you build the baseline from your won and lost deals?

Take every deal that reached your pilot stage and has since closed, as won, lost or no decision, and measure what happened at each later stage. Of the 11 cited pages we read, 3 tell sellers to use their own deal history and only 1 shows how.^↗ Our audit reads 24 months of closed deals, because MES cycles are long enough that a single year holds too few.^↗

Step 3 is the test of an exit criterion. If deals where the plant controller was engaged before the pilot ended win no more often than the rest, the criterion is paperwork, so replace it. If they win clearly more often, keep it in the stage definition and give deals without it the lower rate.

Keep the counts visible while they are small. For example, with 9 closed deals behind a stage, a single deal moves the win rate by 11 points, so show “4 of 9” rather than 44%. Then hold every forecast against the baseline: record the forecast in week 4 of each quarter and set it beside what closed.^↗ Judge the result over several quarters. With 20 deals of €75k, each closing at 30%, chance alone swings a quarter by about a third (Illustrative).^↗ We showed that arithmetic in why a healthy pipeline still misses the forecast.

Stages that demand evidence fix a manufacturing forecast

We think the fix for a manufacturing forecast is stage definitions that demand evidence, and no benchmark can stand in for them. A median cycle describes other vendors’ deals. It can’t tell you which of your pilots has a controller engaged, or which capital request has a meeting date.

Low confidence in forecasts is common. In a Gartner survey published in 2020, 45% of sales leaders and sellers had high confidence in their organisation’s forecasting accuracy.^↗ We read a figure that low as a sign of how forecasts are built, more than of a few weak reps. Coaching still helps a rep run a better pilot. But a forecast can only hold what its stages require, so we would change the stages first.

We don’t ask for every seat in the buying panel. We ask for the ones that can stop the capital request: the operations owner, finance and the approver at the deal’s amount. A deal with those seats engaged is a different deal from one with an enthusiastic plant engineer and nobody else.

Our last reason is the baseline. Without one, any forecast can be called better. Comparing each quarter with your own past deals is, in our view, the minimum bar for saying the forecast improved. We don’t claim stage definitions alone make a forecast accurate. We think they show its errors early enough to act on.