The idea usually arrives on a busy Tuesday. The 6pm is full again, four names are still on the waitlist when class starts, and there is a space across town with a sign in the window. A full room feels like proof, but it proves demand at one hour in one room. A second location copies the whole business behind that room, including the jobs only you do and the gaps nobody has fixed.
So before the lease, run four tests on your own numbers rather than anyone's rule of thumb. One invented studio runs through all four: Alturra Studios, 240 members at $169 a month in one room of 20 spots, whose owner, Sam, is weighing a second site this April. Swap in your own figures as you go. This is a method, not financial advice, so take your version to your accountant before you sign.
The second location in one minute
- Busy is not ready. A waitlist proves demand at one hour; a second site copies the whole business behind it, including the parts that only work because you are there.
- Four tests, your own numbers. Six months of retention, runway on the slow case, a manager bench, and the names your prime classes turn away.
- A failed test is a date, not a no. Each one names its fix. Fix it, run all four again, and give each site its own row from the day the doors open.
| Test | What you count | Alturra in April | Passes when |
|---|---|---|---|
| 1. Retention trend | monthly retention for six months, one formula | 94.5% on average; January's dip explained | steady or rising, every dip explained |
| 2. Runway | savings against the slow case's deepest point, after the flagship's surplus | $48,000 against $54,000 | savings cover it, with a cushion you set |
| 3. Manager bench | hours of flagship work only you can do | 22 hours; she will have 10 | it fits, and two weeks away prove it |
| 4. Turned away | names still waiting when prime classes start, four weeks | 48, at four slots, every week | most weeks, with no prime hour left in the room |
When is a studio ready for a second location?
When four of its own numbers agree, not when the room feels full. Each test checks one thing a second site will either copy or strain:
- Retention is copied. The new site inherits whatever keeps your members, including any habit of yours that stands in for a system.
- Cash is strained. Rent and payroll arrive in full from the first month; members arrive over quarters.
- Your time is strained most. The new site takes most of your week, so the flagship needs someone who already runs it.
- Demand is the reason. People your flagship cannot serve at the hours they want are the case for a second room.
Is your retention steady enough to copy?
It is if six months of it, counted the same way every month, hold steady or rise, and you can explain every dip. One good quarter is not a system. Alturra works it out the way the retention rate formula sets it out: members at the end of the month, minus the people who joined during it, divided by the members at the start, with a short freeze counted as a member under its written rule.
From October to March it kept 94.5% of its members a month on average. January dipped to 93.4%, fifteen cancellations where eleven to thirteen is usual: the autumn's quiet fades, arriving on schedule in the new year. February and March came back to 95.3% and 95.0%.
Two cautions before you call that a pass. At 240 members one cancellation moves the rate about 0.4 points, so read all six months, never the latest one. And read the cohort row, because a new site is all new members for its first year: 27 of the 40 people who joined Alturra from October to December were still members on day 90, 67.5%, where three months at 94.5% would keep about 84% (0.945 × 0.945 × 0.945). That is one more reason to test the slow case below. The churn rate calculator turns your own six-month average into a year.
One question hides in this test: is the rate steady because something notices the regulars who drift, or because you do? If it is you, it will move the month you start spending your days across town, and the third test will find it.
How much runway does a second location need?
Enough savings to carry the new site through the deepest point of its losses in a slow case, after what the flagship can send each month. The burn-rate formula starts the sum; adding it up month by month finishes it.
Start with what the flagship can spare. From January to March, Alturra netted $12,500, about $10,400 and $13,000 before the owner's pay, about $12,000 a month; the chart of accounts piece has its books. Take off what Sam pays herself, $6,000, and anything else that comes out of profit first, such as loan payments, of which Alturra has none. The $6,000 left is what the flagship can send across town each month. Savings, counted after the build-out, deposits and opening costs, are $48,000.
Now the new site. Rent, a manager, the desk and the teaching schedule cost $30,000 a month from the first month, while members arrive over quarters. At $150 a member, founding rates included, it breaks even at 200. The plan: the presale sells those 200 before the doors open, and the site adds 20 a month after that. The slow case, the one to test, sells 80 and adds 10 a month. It loses $18,000 in its first month, $1,500 less each month after, and breaks even in month 13.
The burn-rate formula is runway = savings ÷ monthly net burn, where net burn is the new site's loss minus what the flagship sends: $48,000 ÷ ($18,000 − $6,000) is four months on the first month's loss. The loss shrinks as members arrive, so add it up month by month.
The slow case needs savings through month eight, until the new site's loss falls to the $6,000 the flagship sends. By then they have paid out $54,000 against the $48,000 Alturra has: $6,000 short before any cushion at all. The plan case needs nothing from savings, which is why it is the wrong case to test. It is built to pass.
Run two more lines before you trust the answer. A cushion: Alturra's rule is that savings must cover the slow case even if the flagship's surplus halves while Sam is across town, and at $3,000 a month the slow case needs $82,500. And the lever: every founding member sold before opening takes $150 off every month of the hole. Open with 120 instead of 80, still adding 10 a month, and savings cover $15,000 instead of $54,000; sell the plan's 200 and there is no hole. The presale playbook sets that target from the day-one break-even and paces it daily.
Who runs the flagship while you are at the new site?
A named person who already does it, proven by a real absence before you sign. Not someone you plan to hire once the lease is done.
Sam spends about 40 hours a week at the flagship and expects the new site to take 30 of them for its first six months, which leaves ten. So list last week's flagship work and mark what nobody else can do today. At Alturra it comes to five jobs and 22 hours. The other 18, her own classes and desk shifts among them, are work other people already do and can take.
Twenty-two hours of work only Sam can do, and ten hours to do it in: a fail, and a useful one, because it names the fix. Maya, the desk lead, takes the schedule and subs from a written cover list, member problems up to a written limit, and the first read of the numbers, leaving Sam a one-hour review. The owner-only list drops to eight hours, inside the ten.
Then prove it with an absence: two weeks away before any lease is signed, with the numbers read once a day on a phone and nothing else. The week-off kit is the procedure, and it doubles as this test. If the handover costs money, a raise for Maya or a new desk shift, it comes off the flagship's $6,000, so run the runway sum again with it.
Is your flagship turning people away?
It is if your prime classes are full with names still waiting most weeks, and the room has no hour left at the times those people can come. Demand you cannot serve is the case for a second room. Demand a second session could serve is not.
Count it slot by slot for four weeks: the names still on the waitlist when each prime class started, each one a time someone was told the class was full. At Alturra's four prime slots that happened every week, 48 times in all.
That is not 48 different people: some of Tuesday's names are the same regulars most weeks, a signal of its own. Look for the cheaper fix first, because a second session costs one instructor and a second site costs a lease; reading fill by slot shows where one could go. Alturra's Saturday 10:30am, added last autumn, is why Saturday's list is the shortest. The weekday evenings have no hour left, and an empty 7am does not help someone who needs the 6am: a waitlist is for an hour, not a class. The waitlist piece reads the queue as demand.
Then check where the demand lives. Count the members who live closer to the new address than to the flagship, from the postal codes on your member list if it keeps them. Alturra has 36. If a third of them move, the flagship loses 12 members and $2,028 a month in dues, and the new site counts 12 members it did not win. That becomes growth only when the spots they free fill from the waitlist, so keep movers out of the slow case.
What if a test fails?
Then the answer is a date, not a no. Each failed test names its own fix, and Alturra's two take six months:
- The bench. From April to June, Maya takes the schedule, the subs, member problems and the first read of the numbers. In July, Sam takes two weeks away.
- The runway. If the flagship keeps sending $6,000 a month, six more months take savings from $48,000 to $84,000, enough for the slow case even with the flagship at half strength, which needs $82,500.
In October, Sam runs all four tests again on fresh numbers. If the space across town will not wait six months, that tells you something about the space, not about the tests.
A second location does not test your idea. It tests your systems, and it grades on a curve you did not set.
What should you have in place before you sign?
All four tests passed on your own numbers, and the reporting rules written down. The first pass takes an afternoon with last month's reports.
What to do this week:
- Work out six months of retention by one formula, joiners left out, and your newest cohort's share still members on day 90.
- Take the flagship's net over its last three closed months, less your pay and anything else that comes out first. That is the surplus.
- Model the new site month by month on the slow case and add up what savings must cover. Run it again with the surplus halved.
- List last week's flagship work, mark what only you can do, and set those hours against the ones you will have.
- For four weeks, count the names still waiting when each prime class starts, and check whether a second session could serve them first.
- Write the home-location rule and the rest of your definitions before the doors open.
What changes the day you open?
Every number you relied on gets a second meaning, because two sites now sit inside every total. Set the rules before opening day, or you will spend the first year restating your own history.
- Each site gets its own row, beside the company total. The total hides a struggling new site for months.
- Each member gets one home location, by a written rule. First visit or most visited, written into your definitions and never switched, or the sites will not add up to the company and two honest reports will disagree.
- Rates are recomputed, never averaged. Sum the counts across the sites, then divide.
- The new site is judged against its own plan and slow case, and against the flagship at the same age, never against the flagship today.
The company's revenue is up $21,000 a month, and its churn has drifted from the flagship's 5.0% to 6.8%, the kind of move an owner puts down to the season. The new site's own row says it lost $9,000 this month and is losing members at twice the flagship's rate, and an average of the two rates would have said 7.5%. The per-site scoreboard piece sets out the three rules that keep a multi-location report honest.
The same rule holds for the funnel. In Xyzios, the Compare board reads each location and the combined figure on the last closed cohort, and recomputes the combined rates from summed counts rather than averaging them.
Where this lives
Who runs the flagship, and whether the space is worth its rent, are decisions no software makes. The numbers around them are where Xyzios, the studio operating system, comes in. It works with Mindbody and is an approved Mariana Tek integration, and your booking platform stays the system of record.
You decide what counts as a member and whether a freeze counts as churn, and the boards follow your rules. Current members whose attendance is slipping against their own pattern land on an at-risk list, with a save note drafted for those at high risk, so noticing the drift does not depend on who is in the building. The Programming board flags the class with the heaviest waitlist, and with the books connected it shows one estimated break-even headcount per location. With QuickBooks Online connected, the P&L your runway sum starts from is in Finance.
For the new site, a location can be planned, in presale or live: a presale site gets its own board, with founding sales against a daily pace line, and joins the company rates on opening day. From then on each location has its own row on a per-location scoreboard with the company total, the Compare board reads them as above, and a location manager can have her own locations' list in the morning email if you switch it on. Nothing reaches a member until you tap approve. There is more on opening a location with Xyzios and on running several.
Whatever you run it in, run the four tests before the lease, not after it. The two-minute studio check shows which of your numbers needs you first.