# Running a time study

> Time one task with the AI and without it, publish the result as your formulas' before and after figures, and what it shows and doesn't.
>
> Source: https://roiva-staging.com/docs/running-a-time-study
> Checked against the product on November 09, 2026

A time study measures how long one task takes the people who do it, with the AI and without it. Its result becomes the figures a formula compares, such as agent minutes per ticket before and after. It's how to measure a baseline that no system records, and how to reconstruct one once the AI is already in use. See [Baselines for an initiative that's already live](https://roiva-staging.com/docs/baselines-for-an-initiative-already-live).

Time studies live under **Method → Time Studies**. An Owner or Admin, or the initiative's owner, can start one.

#### Setting one up

1. Choose **New Time Study**. Pick the initiative and the metric it measures. Only metrics the initiative's formulas read are offered, or agent minutes per ticket when its help desk prices deflected tickets from a study. Those are metrics counted in minutes, and a few of Roiva's counted in hours of one person's work on one unit done in a day, such as engineer hours per deployment or KYC review hours per account. People still time those in minutes, as a number such as 90 for an hour and a half, at most eight hours in one timing, and the result is published in hours. An hours metric that totals a week or a month, adds up several people's hours, covers a whole cycle, or counts a unit worked over several days, such as a bid, isn't offered: a study times one task in one day. For one of those counted per unit, take it from time tracking instead: the hours logged on that work ÷ the units finished, over a period before the AI and one after, entered as its baseline and a reading. See [Metrics, baselines, and readings](https://roiva-staging.com/docs/metrics-baselines-and-observations).
2. Describe the task as participants will read it. Say what's in and what's out: "One ticket, from first reading it to marking it solved. Count only the minutes you spend on it, including checking what the AI drafted."
3. Choose what it times: with the AI and without it, or one of the two. With both, Roiva draws each participant's days without the AI at random, so nobody chooses which days go into which arm.
4. Set the window, the time zone its days run in, the confidence of its interval (90% or 95%), and how long its result stands before the task is due to be timed again (3, 6 or 12 months).
5. Set the precision you want and how much the task varies. The study's page then says how many timings each arm needs, at least 30. Once timings arrive, it says how many more reach that precision. Aim for at least five people in each arm. Three is the fewest a result can be published from, and with so few the interval understates how much people differ. From the third day, if an arm has fewer timings than its share of that number for the days gone, the study shows in the **Inbox** of whoever manages it, once a week while it stays behind.

#### Who times the task

Add the people who do the work.

- **A member of the account** records in the app, signed in. Their Inbox shows the study each day, with that day's arm.
- **Anyone else** gets a link of their own, under a label such as "Agent A" rather than their name. Roiva never emails the link: copy it and send it yourself. It's shown once, stored only in a form that can't be turned back into the link, and stops working a week after the window ends, or as soon as you revoke it or issue a new one. The page it opens shows the task, the study's name and dates and the account's name, and nothing else in the account.

Each person times one task at a time with the page's timer, pausing it when they step away, or types the minutes. The study says which arm each day is. They can't change a timing once it's recorded. If one is wrong, tell the study's owner, who can leave it out of the result with a reason from the study's **Timings** tab; it stays on the record. The tab is the owner's alone, since it lists each person's minutes one by one; if the owner leaves the account, it passes to the account's Owners and Admins and the initiative's owner.

The timer survives a reload, a closed tab or a second tab. The browser keeps the timing in progress, on that computer only, and records it on the day it started, even after midnight. If the computer sleeps or the page is closed while the timer runs, it pauses where it stopped and asks whether the time since was spent on the task. A timing that can no longer be recorded, nine hours after its page opened or once its day is past yesterday, is cleared, and the page says so. **Pop out** opens the timer in a small window, and Alt+S starts or pauses it while the page is in front. **Not you?** on a link's page also clears any timing that browser kept.

#### Reminders after meetings

Timings get lost when people forget the timer. If the task usually follows a meeting, such as writing it up, whoever manages the study can have Roiva email a reminder when a meeting ends. Under **Reminders after meetings** on the study, choose the account's **Google Workspace** or **Microsoft 365** connection, the shortest meeting worth a reminder, at most how many a day, and the hours to send them in, in the study's time zone.

- **Each person chooses.** A participant who's a member of the account turns reminders on, or off, on their recording page, and every reminder has a link that stops them. Someone outside the account has no email in Roiva, so isn't reminded.
- **What Roiva reads.** Only when each person's calendar shows them busy, through the connection, as any colleague sees when scheduling a meeting: never a meeting's title or who's in it. Microsoft 365 can send a title along for a calendar shared in detail; Roiva drops it unread. Of the calendar, Roiva keeps a digest of when a reminded meeting ended, so the same meeting isn't reminded of twice, and the day each reminder went. Roiva's mail log, like every email it sends, keeps when each reminder was sent, which is minutes after a meeting ended. Someone whose calendar isn't shared with the connection, or that Roiva can't read for another reason, is told on their recording page.
- **When.** Checked every ten minutes, a reminder goes when a busy stretch of at least the chosen length has just ended and the next ten minutes are free. Back-to-back meetings count as one, and anything over four hours isn't a meeting. None goes past the day's maximum, once someone has recorded the most timings a person can in a day, or once every arm has as many timings as the study asks for. Someone in two of an account's studies is reminded of a meeting once, and someone Roiva has deactivated isn't reminded.
- **Both arms alike.** Reminders go out the same on days with and without the AI, so they don't tip one arm more timings. The email says which arm the day is.

#### Importing timings from another tool

If your team already tracks time in Zendesk, Freshdesk, Harvest, Toggl or Jira, the study's owner can choose **Import timings** on the study and upload a CSV of its export. It needs a header row naming these columns, one task to a row:

- **participant**: a member's email, or a label of up to 60 characters for someone outside the account. Many tools export full names: replace them with labels first if you don't want names kept.
- **arm**: with the AI, or without it.
- **minutes**: hands-on minutes, such as 12, 12.5 or 12:30.
- **date**: YYYY-MM-DD, inside the study's window.
- **task** (optional): a ticket or task ID.

If any row can't be read, nothing is imported, and the rows are listed. A row already imported is skipped, so importing the same file twice records it once. Roiva doesn't keep the file.

A tracker workbook's **Time studies** and **Timings** sheets bring in whole studies instead, each new, with their timings marked imported in the same way: importing the workbook again creates them again. Its minutes are numbers, such as 12.5. See [Importing a tracker workbook](https://roiva-staging.com/docs/importing-a-tracker-workbook).

Roiva didn't draw the days of imported timings, so whoever recorded them chose which went into which arm. The result says how many timings were imported, and a study with any publishes its without-AI figure as a sample, never as measured before go-live.

#### Checking days without the AI

A without-AI figure is only as good as the days really were without it. If the account connects **Claude Enterprise**, an Owner or Admin who manages the study can check them: under **Days without the AI** on the study, choose the connection to check against. Reading people's use of Claude by day is an Owner's or Admin's, as it is on the connection itself, so the check runs only while whoever turned it on still is one.

- **What it checks.** For each participant who's a member of the account, on each day the study assigned them without the AI after the check was turned on and they were added to the study, whether Claude Enterprise's activity shows them using any Claude product that day: chat, Claude Code, Cowork or another. They're matched by the email they sign in to Roiva with. Someone outside the account has no email in Roiva, so isn't checked, and someone who leaves the account isn't checked after they go.
- **When.** Each night, two days after a day ends, once Anthropic has reported it, and when the owner chooses **Check now**. It reads no day from before it was turned on, since participants are told only while it's on, so turn it on before the study starts to check every day. It stops once the study is published or archived. If whoever turned it on stops being an Owner or Admin, it pauses and says so until an Owner or Admin saves it again.
- **What it can't say.** Anthropic reports whole days in UTC, so it says whether someone used Claude that day, not on which task or at what time. The study's day is read as the UTC day of the same date, so use the evening before, in a zone behind UTC, counts on the day. A UTC day that began before someone was added isn't read for them. A member it never finds in Claude Enterprise's activity during the study, perhaps under another email or without a seat, is listed as not found, and their days count as unchecked.
- **What happens.** The owner sees who used Claude on which day, with the products, and the timings recorded on those days are marked. Nothing is excluded on its own: the owner can leave those timings out with one click, each with the reason, or keep them if the use had nothing to do with the task.
- **Who sees it.** The study's owner, as with timings (if the owner leaves the account, whoever manages the study), and the account's export. Participants are told on their recording page while the check is on. The study's owner, or an Owner or Admin who manages it, can turn it off, which removes what it found, even once the study is published. So does choosing another connection.

#### The result

The study's owner sees the result as timings arrive. Whoever manages the study sees it once it's closed with enough people to publish, and everyone in the account sees it once it's published.

For each arm it shows:

- the mean time per task, with an interval, in minutes or, for a metric counted in hours, in hours;
- how many timings it rests on, from how many people;
- the median, when a few very long tasks pull the mean away from it.

The interval is worked out by person. Twenty timings from one fast person are one person's pace, so the interval narrows only as more people are timed. With both arms, it also shows the difference between them and its interval. When the initiative has a formula comparing the metric before and after, and last month has every figure it reads, it also shows what that range is worth a month at last month's volume and rates, and whether that's cash or freed capacity.

The figure published is the mean, not the median: a formula multiplies the time per task by volume, and only the mean times the volume is the total.

#### Publishing

Close the study, then choose **Publish Result** at the top of its page. Each arm needs timings from at least three people. A dismissed initiative is measured no more: dismissing one closes its studies that are still taking timings, and their results can't be published.

- The without-AI mean becomes the initiative's baseline for the metric. It's labeled measured (a time study before go-live) only when nothing in the study used the AI. Otherwise it's a without-AI sample, a reconstructed baseline.
- The with-AI mean becomes a reading of the metric.

Both stand from the month the window started until the study's re-study date. Roiva recalculates each month's figures from them; anything already approved stays as it was approved.

A study of agent minutes per ticket also prices the initiative's deflected tickets, from the help desk's next sync, if they were on Roiva's default price: its without-AI minutes multiplied by your labor rate. See [How deflection value is booked](https://roiva-staging.com/docs/how-deflection-value-is-booked). A published study takes no more exclusions, and its figures can't be edited: to change them, time the task again in a new study.

Whoever published the result, the study's owner, whoever left timings out, whoever imported timings and every member who timed tasks for it count as authors of the figures, as whoever types a reading does. Your approval settings then decide whether they may approve value calculated from them. See [Approving value and initiatives](https://roiva-staging.com/docs/approving-value-and-initiatives).

#### Timing it again

From a month before a study's re-study date, whoever manages it (its owner, the initiative's owner, and the account's Owners and Admins) sees it in their **Inbox**. After that date, value calculated from its figures is marked **Re-study due** and held for a person to approve. Publishing a new study of the same task ends that. The months before the new study's window keep the figures they were calculated from.

On a closed or published study, **Start a new study** opens the new-study form with its initiative, metric, task and plan settings filled in, over a window of the same length starting today. Nothing it collected comes with it.

#### Archiving and deleting a study

A study that is closed or published can be archived from the more-actions menu (the three dots), which takes it out of the Time Studies list and the Inbox. Archiving changes nothing else: a published study's figures keep standing and its timings stay. **Show archived** on the list brings archived studies back, and **List again** on one restores it.

An Owner or Admin can delete a study that is closed and was never published, by typing its name to confirm. Deleting removes its participants and every timing, and can't be undone. The audit log keeps that the study was deleted, who deleted it and how many timings went, with counts only and no participant's name or label. Earlier audit entries about adding participants or issuing their links keep the labels they were given. Exports already downloaded keep their copy, and database backups keep theirs for 30 days, used only to restore Roiva. A published study can't be deleted, because the figures it published and the value calculated from them name it: archive it, and time the task again in a new study.

#### What a time study shows, and what it doesn't

It shows how long the timed tasks took the people timed, in that window, with an interval for how much it could differ had other people been timed.

It doesn't show:

- that the tasks timed stand for every task;
- that people work at the same pace when they aren't being timed;
- that typed minutes are exact;
- what any minutes freed were used for;
- that the AI caused a change in volume or outcomes.

It's a measured estimate of the time a task takes, not an experiment. A measurement charter can name the study a baseline came from. See [Measurement charters](https://roiva-staging.com/docs/measurement-charters).

#### Related

- [Baselines for an initiative that's already live](https://roiva-staging.com/docs/baselines-for-an-initiative-already-live): the other ways to set a baseline
- [Metrics, baselines, and readings](https://roiva-staging.com/docs/metrics-baselines-and-observations): how formulas read figures
- [Aggregate Figures Only: a note for your risk team](https://roiva-staging.com/docs/aggregate-figures-only): what a study keeps about a person outside the account
