Trying Out New Site Software? Decide What Success Looks Like Before You Start
On day 30 of a pilot, the vendor shows a chart of logins and the site team says the app was fine, and then the developer turns to you and asks whether it worked. You can't answer, because nobody agreed at the start what working would look like, so the decision slides into another month or gets quietly dropped. If it is dropped, the record that might have settled next year's payment dispute never reaches your other sites, and if it is waved through, your name sits on a tool the site teams stop opening within a quarter.
A construction technology pilot should prove one thing you agreed before it started, measured on one live site, with a yes-or-no rule for day 30. For progress monitoring, the measure is usually whether the dated record let you sign off progress without a site visit, which only happens if the site team keeps capturing. The way construction technology is now being sold shows how to avoid that meeting.
Why one number now decides
At the end of October, a construction-robotics maker announced that its machines had completed 300,000 tasks on utility-scale solar projects, and the same report explained that its customers pay per task completed (Robotics & Automation News, 30 October 2026). Paying per task means buyer and vendor fix the measure in the contract, so the result needs no interpretation at the end.
Most construction technology buying hasn't reached that point. In a survey of more than 100 construction and development executives about their AI spending, published in October, 37% said they judge its returns mainly on financial metrics, 29% on operational measures and about a quarter on usage or adoption, while 6% don't evaluate returns at all (BRG, 13 October 2026). When you, the contractor and the vendor each hold a different one of those yardsticks, the same 30 days can look like a success to one of you and a failure to another.
Agree the measure before day one
Start from the decision the tool is meant to change, because a measure that isn't tied to a decision gets argued about at the end. For an owner's representative, progress monitoring usually exists to change one of a few decisions: signing off progress for a payment, deciding whether you need to travel to site, or settling a question about work that is now covered. Choose one for the pilot, since 30 days rarely holds enough of each event to prove several at once.
Then write down how that decision is made today, which becomes your baseline, and agree who will count the result on day 30, putting the measure and the day-30 rule in writing with the contractor and the vendor before the first capture. A trade-press guide to running tech pilots puts the same point in one line: "Know what you're looking for, have a goal in mind" (Construction Dive, 6 November 2024).
Measures that fit a 30-day progress-monitoring pilot
| The decision you want to change | The measure to agree before day one | How you check it on day 30 |
|---|---|---|
| Signing off progress for a payment | Sign-offs made from the dated record, without a visit | Count sign-offs and visits in the window, against last month |
| Whether you need to travel to site | Visits made only because the record raised a question | List each visit with the reason for it |
| Settling questions about covered work | Questions in progress meetings answered by opening two dates | Note each question and whether the record answered it |
| Whether the record will exist at all | Walks done against walks planned, on the agreed route and day | Compare the capture dates with the plan |
Track the last row whatever your main measure is, because a record that was never captured can't support any decision.
Choose one live site
Run the pilot on one live site with months of work left, because a site that barely changes in 30 days gives you nothing to compare. Pick one where a decision you care about falls inside the window, such as a payment cut-off or a ceiling close-in, so the record is tested against something real. Agree with the contractor which route the site team will walk and on which day each week, since the measure depends on that habit more than on any software. Your best-run site is a poor choice, because a pilot that works only where everything already works tells you little about the rest of the portfolio.
Decide what "worked" means after 30 days
On a programme with monthly valuations, 30 days holds four weekly walks and one payment cut-off, which is enough to see whether the capture habit holds and whether the record answered your question. Write the result as yes, no or extend, and decide in advance what earns an extension, such as a walk missed for reasons outside the team's control, so you walk into the day-30 meeting already knowing how the answer will be read. A yes means the next test is a second site of a different type, because that is where the decision to roll out is really made.
- Thu 1 OctMeasure agreedOne decision, a baseline and the day-30 rule, in writing
- Thu 8 OctWeek 1 walkThe first dated record of the agreed routeRoute A
- 15 and 22 OctWeeks 2 and 3Same route, same day, so any two dates line upRoute A
- 29 and 30 OctWeek 4 and cut-offSign off from the record, or log the visit and its reasonRoute A
- Sat 31 OctDay 30Yes, no or extend, read against the rule
Where SpatialSense fits
SpatialSense has a 30-day free trial of the full platform with no card, which matches the window above, and nothing is charged when it ends. Every plan has unlimited users, so the contractor's site team, your own staff and the developer can all join the pilot, and support is one-to-one with a person at Kodifly, 24 hours a day, 7 days a week, which matters while the pilot clock is running.
During the trial the site team uploads 360° panoramas of each walk in the browser, or records walks with the Trace app, which is set up with you when you book a demo. Each walk is dated, any two dates open side by side from one locked viewpoint, and from two dated walks SpatialSense writes a station-by-station progress report tied to the images, as a PDF. Share links open without a login and can be revoked, so the developer can check the record without an account.
On a thirty-week new build in Islamabad, the people already walking the site made seven dated captures in twenty-four days with no survey crew, and the client's architect reviews progress weekly from Europe without flying in; our post on weekly sign-off without a site visit shows how that week runs. On a residential tower in Lahore, two walks 48 hours apart were enough for the AI progress report to list what changed, station by station.
SpatialSense is a weaker fit if your pass mark is a two-way sync with an existing project-management system, because no two-way sync is published. If the decision you want to change is a measured one, such as excavated volume, plan the pilot around laser scans, because a 360° walk shows the work without measuring it.
Capture. Understand. Take Action.
See SpatialSense on your project
Write down the one measure and the day-30 rule, agree the site and the walk day with the contractor, then start the 30-day free trial with no card. The trial page lists what the 30 days include, or you can book a demo and we set up Trace for your site team.
FAQ
What should you measure in a construction software trial?
The decision the software is meant to change, such as sign-offs made from the record without a site visit, plus whether the team kept capturing on the agreed day, with a baseline taken before the trial starts.
How long should a construction technology pilot last?
Long enough to cover the decision you are testing at least once. For progress monitoring, 30 days usually holds four weekly walks and one payment cut-off, which shows whether the capture habit holds.
How do you measure the ROI of construction technology?
Firms use different yardsticks, from financial returns to operational measures and adoption. In a pilot, count the decision the tool changes, such as visits avoided or questions settled from the record, and put a money value on it afterwards.
Who should take part in a progress-monitoring pilot?
The site team who will capture, the person who signs off progress and whoever decides on roll-out, because each of them has to trust the same record. Agree the measure with all of them before the first walk.