help & docs
most people come here when something doesn't make sense. we'll be plain about it.
team health
what is a persona?
a persona is a recognizable pattern of feeling about a domain. it's not a label for a person — it's a name for a shared shape across responses.
how anonymity works
an answer is stored as the check-in it belongs to and your three choices, plus the band you pick if your team uses bands. no name, no email, no user id, no time of its own. a separate "did you answer" ledger exists so you can see your own history. only you see it, and it is never joined to your answers. if your team uses bands (like tenure), the band you pick goes on your answer, and a band is only shown once 5 people share it. like every database record, an answer has a creation time in the database's own metadata. we never read it, export it or show it. nobody in the product can link an answer back to you: not your manager, not jollygig. we need 5 people to respond before a pattern shows up — below that, we tell you we're still building, instead of fudging it. managers don't answer their own team's check-ins (they're the audience), so the 5 counts your reports only.
what are bands?
we group responses by tenure, role level, and location so we can spot when a pattern is concentrated in one slice of the team. each band needs its own 5 responses before it shows up. you'll never see who is in a band — just whether a pattern is more present there.
confidence levels
strong signal (at least 5 people answered, and at least 70% of the team), early signal (at least 5 answered, but under 70% of the team), still building (fewer than 5 answered, so no pattern is shown).
lifecycle states
active now (a pattern that keeps showing up across your recent check-ins), emerging (early signal worth watching), not detected (quiet for now). patterns are worked out when a check-in closes, from the last 6 check-ins with the newest counting most, and they need to clearly rise or fall before the label changes, so it doesn't flip back and forth.
writing action plans
address conditions and culture, not individuals. small structural moves beat one-off conversations. plans are private to the manager.
joining a company or moving a team
if your company already uses jollygig, an admin can invite you to its organization, and you accept before anything changes. if you run a team of your own, you can bring it along: the team and its people move into the company's organization and count in its plan, nothing about the team changes, and you stay its manager. your old solo plan stops: a free trial ends the day the team moves, a paid plan runs to the end of the period you already paid for and is never charged again. from then on the organization's owner handles billing; you see a line saying so on the billing page. you can also leave the team where it is and stay in the company's organization as a member.
survey cadence
default is weekly. each check-in opens around 05:00 utc on the day it's due and stays open for 5 days, or 7, or until the next one is due. you choose in team settings. members can pause for a quiet week.
scheduling
workspaces
a workspace is one project: its activities, dependencies, calendar, forecasts and people. a scheduling plan covers up to 10. the owner adds people as editors, who can change the plan, or viewers, who can read it. you can draw the work on a canvas or move cards on a board; they're the same activities either way.
linking activities
a link says one activity has to finish before another can start. on the canvas, drag from the small circle at a card's right edge onto another card. or open a card (details) and tick the activities it depends on. the board's add form has the same list, so a new activity can say what it waits on from the start. to remove a link, click it on the canvas or untick it in the card. activities with no links are assumed to run at the same time, which makes a plan look faster than it is, so link what really waits on what. the tidy up button lays the cards out left to right in the order the work runs, and a linked plan is laid out that way for you when you open it with its cards still stacked in one column.
three-point estimates
every activity has an optimistic, a most likely and a pessimistic duration, in workdays. think best plausible and worst plausible, not best or worst imaginable. the three have to run in order, smallest to largest.
distributions
the three numbers become a shape. beta is the default: it leans toward the most likely value, and suits work you have history for. triangular peaks sharply at the most likely value, a better fit when it's more guess than data. lognormal has a long tail, for work that rarely but really runs very long, and needs a best case above zero. each activity has its own, and mixing is fine.
activities that run long together
for any two activities you can tick what they share: the same person or team, the same unknown, the same technical area, or an estimate anchored on the other. the forecast then draws their durations together, so the long end stays honest. you never enter a coefficient. if the ticks can't all be true at once, we say we softened a couple so they fit.
calendars and holidays
each workspace has its own calendar: which weekdays are workdays, which dates are holidays, and partial days at a quarter, half or three quarters of capacity. durations count workdays on it. us federal holiday presets for 2026 and 2027 are one click.
reading the results
the forecast panel shows when the plan might finish: the p50 (half the runs finish by then), the p75 and the p95, a chart of how the runs spread, what's moving the date, and which path sets the finish. activities are described in words, like 'on the critical path sometimes', never as a score. set a committed date and you also see the chance of making it. a committed date on a day off (a sunday, a holiday) is counted as the last workday before it, so a date promised by sunday means done by the end of friday, and the forecast says so next to the chance. chances are rounded to the nearest 5% below 90, and never shown as a flat 0 or 100.
logging actuals
move a card to in progress and its start is recorded; move it to done and its finish is. for work in progress you can re-estimate what's left. finished work counts as fixed, the forecast re-runs on the rest from a 'forecast from' date (today unless you set one: work that hasn't started can't begin before it), and each run is saved so you can watch the chance of your date over time.
the strain note
if a workspace is linked to a team in team health and that team is showing signs of strain, the forecast panel shows a quiet note: estimates made under strain tend to run optimistic, so widen the pessimistic ones by 10%? one click applies it when the forecast runs. your own estimates are never changed, the widening is shown next to the results, and you can reset it any time. it never names a person or a persona.
billing
the free trial
30 days, no card, for whichever products you chose at signup. about a week before it ends we email a reminder. add a card any time: the plan you chose then starts when the trial ends, and the card is first charged then. add none, and the plan ends.
cards and invoices
the owner adds or changes a card on the billing page, which opens stripe's customer portal. invoices are there too, and stripe emails a receipt after each charge. jollygig never sees or stores a full card number, only the brand, last four digits and expiry.
changing your plan
the owner can add or drop a product, move between starter and growth, or switch between monthly and annual on the billing page. changes apply right away: the extra cost for the rest of the period is charged at once, and a lower price is credited to your account. with both products, scheduling is 20% off, automatically.
canceling
the owner can cancel from the billing page. it takes effect at the end of the period you've paid for, everything keeps working until then, and you can reactivate before it ends. nothing is deleted when a plan ends. the organization turns read-only and stays that way for as long as it exists: everyone can still sign in, read and export, and the owner can resubscribe any time. the owner can delete teams and workspaces whenever they like, and once the owner is the only person left in the organization, deleting their account removes the organization and everything in it immediately. to close an organization with other people in it, the owner can use the close this organization button on the organization page, which ends the subscription and deletes its teams and workspaces at once; or remove or transfer the teams first, then delete the account. database backups are kept for 14 days, so deleted data has left those copies within 14 days. anonymous answers are still removed once they are older than each team's retention period (24 months unless the team chose another).
read-only, not locked out
when a trial or plan ends, or a payment still hasn't gone through after 14 days of retries, the organization turns read-only. you can sign in, read everything and export it. nothing can be edited and new check-ins stop opening. nothing is deleted. a banner says why, and an owner or admin can subscribe again or update the card from it.
refunds
we don't refund automatically. charged by mistake, or canceled in the first 14 days of paying without really using it? write to help@jollygig.com and a person will sort it out.
frequently asked
team health
scheduling
billing
still unclear? write to help@jollygig.com and a person will answer.