Waiting List vs Open Launch: How Monzo Controlled Its Early Growth
Short answer: Use a waiting list when admitting every buyer would exceed safe fulfilment, support or compliance capacity. Open the launch when each extra customer can receive the promised outcome without a rising failure rate. Release customers in measured cohorts, compare capacity with activation and support demand each week, and remove the gate once it protects nothing except artificial scarcity.
Waiting lists are often described as marketing theatre. Scarcity can attract attention, but attention is the weakest reason to delay a willing customer. A useful queue is an operating control: it keeps demand below the level at which service quality, cash or trust breaks.
Monzo began in 2015 with about 5,000 alpha prepaid cards before widening access. Its published chronology says the 2016 crowdfunding round attracted £1 million in 96 seconds and that roughly 10,000 customers later suggested replacement names when the business moved from Mondo to Monzo. Monzo's early chronology records those milestones. The bank describes its Golden Ticket referral scheme as an early product-led growth mechanism through which existing users gave friends access. Monzo explains the role of Golden Tickets here.
Those later outcomes do not prove the queue caused growth. They show how controlled access can combine learning, referrals and operational pacing when the product cannot safely admit everyone at once.
Use the Controlled-Access Growth Valve
The Controlled-Access Growth Valve sets admissions from evidence rather than excitement.
| Valve reading | Measure | Open further when | |---|---|---| | Delivery | Orders or accounts completed correctly | Backlog stays within the promised window | | Activation | Admitted customers reach first value | Cohort activation meets your threshold | | Support | Cases and minutes per new customer | Team capacity remains below 80% | | Reliability | Failures, refunds and incidents | Rate is stable or falling | | Cash | Cash needed before revenue arrives | Downside case remains funded | | Trust | Complaints and unresolved harm | Serious issues are closed before expansion |
My position is that you should never use a waiting list merely to look popular. It asks a ready customer to wait and conceals whether the offer converts. Use it only when each cohort produces learning or protects a measurable constraint. Otherwise, open the launch and let real buying behaviour judge the proposition.
Find the constraint the queue protects
Name one binding limit. It might be manufacturing output, appointment hours, onboarding staff, delivery slots, supplier stock, fraud review or capital. "We want a controlled launch" is not a limit.
Calculate weekly safe admissions:
safe admissions = available weekly capacity divided by capacity required per activated customer
If you have 30 support hours and a new account uses an average of 45 minutes in its first week, support permits 40 activations. If only 80% of invited people activate, you could invite 50: 40 ÷ 0.80 = 50. Add a safety margin when failure would harm customers.
Regulated financial products require far more than support arithmetic. Authorisation, safeguarding, capital, identity checks, complaints, data protection and consumer duties vary by country and product. A founder must use the applicable regulator's current rules and qualified legal and compliance advisers. Monzo's path is a case study, not a shortcut into banking.
Build a queue that measures intent
A sign-up should capture only what you need and tell people what happens next. Record acquisition source, date joined, relevant eligibility and the cohort invited. Do not ask for sensitive data before you have a lawful need and suitable controls.
The queue should distinguish curiosity from intent. An email address after a viral post is weak. Confirming eligibility, choosing a delivery window or placing a refundable deposit is stronger, provided the commitment is appropriate and lawful.
Do not report the waiting-list total without the invitation and activation rates. Ten thousand names with 5% activation can be less useful than 500 people with 60% activation. See [pre-orders versus a waiting list](/business-ideas/pre-orders-versus-a-waiting-list/) when cash commitment, rather than capacity pacing, is the central issue.
Release cohorts, not random drips
Admit customers in identifiable weekly groups. Keep the offer and onboarding process stable long enough to compare them. For each cohort measure:
- invitations sent and accepted
- customers reaching first value
- time to first value
- support cases and minutes
- failures, refunds and complaints
- referrals that activate
A referral pass can improve the quality of the next cohort because an existing customer provides context and trust. It can also distort the sample towards similar networks. Keep a control cohort from the ordinary queue so you can compare activation and support.
Golden Tickets are interesting because access itself was shareable. The useful lesson is not to copy the name. It is to connect growth to a customer who has experienced the product while retaining an admission limit.
Worked example: PocketCurrent Banking
PocketCurrent is a fictional budgeting account preparing a limited pilot. It has 120 support hours available each week. Existing tests show that an activated customer needs 50 minutes of first-week support. The team keeps 20% of capacity for incidents and existing users.
Available onboarding capacity is 120 x 80% = 96 hours, or 5,760 minutes. Maximum weekly activations are 5,760 ÷ 50 = 115.2, rounded down to 115.
The waiting list activates at 65%, so the maximum invitations are 115 ÷ 0.65 = 176.9, rounded down to 176. PocketCurrent starts lower at 140 invitations.
| Weekly result | Amount | |---|---:| | Invitations | 140 | | Activated at 65% | 91 | | Support minutes at 50 each | 4,550 | | Support hours used | 75.8 | | Capacity reserved for onboarding | 96 hours | | Remaining onboarding buffer | 20.2 hours |
If support costs £24 an hour, first-week support costs exactly 4,550 ÷ 60 x £24 = £1,820, or £20 per activated customer. If week-two cases fall and activation remains stable, the next cohort can expand. If incidents consume the 20.2-hour buffer, admissions stay flat. These figures are illustrative and do not represent the costs or regulatory requirements of a real bank.
Know when the waiting list has become the problem
A long wait creates stale demand, repeated enquiries and mistrust. Measure the proportion of invited people who still activate by time spent waiting. If activation falls sharply after four weeks, either shorten the queue or communicate a realistic date.
Open access when capacity is no longer binding, recent cohorts activate consistently and service failure does not rise with volume. You can retain eligibility or geographic limits that are genuinely required, but remove decorative friction.
Do not confuse an open launch with unlimited promotion. You can open purchasing while controlling advertising spend, geographic coverage or delivery slots. The customer should encounter a truthful availability rule, not an unexplained queue.
Set the first valve this week
Within two days, identify the one resource most likely to fail under demand. Measure how much one activated customer consumes. Set a cohort size with a 20% safety buffer, define activation and failure thresholds, and invite the first group. Review the numbers seven days later. Increase admissions by no more than 25% when the buffer survives; hold or reduce them when service deteriorates. After three stable cohorts with unused capacity, test open access for a bounded period and compare results.
Related guides
Frequently asked questions
Does a waiting list create demand?
Not by itself. A queue can signal scarcity and collect interest, but it does not create the underlying problem or willingness to buy. Measure how many people accept an invitation, activate and remain after waiting. If the list grows only because entry is fashionable, conversion will expose it. Use the queue to pace a real constraint and learn from cohorts. For demand validation, ask for an appropriate commitment and compare behaviour with [the amount of demand you actually need](/market-research/how-much-demand-is-enough/), rather than presenting sign-ups as customers.
How often should I admit a new cohort?
Weekly cohorts suit many digital or service launches because they allow enough time to observe activation and early support while preserving momentum. Physical products may need a production or delivery cycle, and complex business services may require monthly cohorts. Choose a period long enough for first value and the most common early failure to occur. Do not expand simply because invitations were accepted. Review the constraint the queue protects, the safety buffer and unresolved complaints before each release. Keep cohort labels so later retention can be traced to launch conditions.
Should people know their exact position in the queue?
Only show a position if it is accurate, meaningful and updated consistently. A number can motivate action, but it also creates expectations about timing that may be false when eligibility, geography or capacity changes. A realistic invitation window is often more useful. Explain what controls admission and send updates when the estimate moves. Do not manipulate positions to manufacture urgency. Data and consumer rules apply to queue communications, so collect minimal information, honour marketing choices and check current local requirements for automated decisions or eligibility screening.
Can I charge a deposit to join the waiting list?
Yes, when the deposit serves a genuine reservation purpose, the terms are clear and you can refund or fulfil reliably. A deposit is stronger demand evidence than a free sign-up, but it creates cash-handling, accounting and consumer obligations. State whether it is refundable, when it becomes payment, what happens if timing changes and how cancellation works. Do not use customer deposits as risk-free finance. Rules vary by jurisdiction and sector, especially for financial services, travel, housing and regulated products. Obtain qualified local advice before taking money.
What if invited customers do not activate?
Separate queue quality from onboarding failure. Contact a sample promptly and ask what stopped them: changed need, unclear value, trust, eligibility, effort or timing. Compare activation by acquisition source and waiting time. If one source supplies low-intent names, reduce it. If suitable customers stall at the same step, fix that step before sending more invitations. Do not compensate for weak activation by flooding the top of the queue. More invitations can hide the problem while increasing support and reputational cost.
When should I abandon the waiting list completely?
Remove it when the protected constraint has spare capacity across at least three representative cohorts, activation and failure rates are stable, and an open test does not degrade service. Also remove it if the queue creates more lost demand and administration than the risk it controls. Keep appointment calendars, stock notices or eligibility checks where they reflect reality, but call them what they are. An honest "next delivery available Tuesday" is better than an invite-only performance. Continue monitoring capacity because open access can expose peaks that cohorts concealed.
Comments (0)