Your app rating is sitting at 2 stars. Downloads have been sliding for a year. There are a few hundred thousand members in the loyalty program, but most have not ordered through the app in months. And more than half of your digital orders still arrive through Uber Eats or DoorDash, who take the commission and keep the customer data.
All of it is fixable, and there is real revenue in fixing it. Which is usually when someone on the executive team asks what a new, high-converting mobile app would cost - and needs a figure for the board before anyone has worked out what the app has to do.
So you ask around. A peer says theirs came in at $300,000. Someone at a QSR conference says well north of a million. Both are telling the truth. They are describing 2 very different pieces of software, and both are called a mobile app.
A mobile app is not a website with a login
It is easy to invest in a mobile app the way you invested in your last website, and to justify it the same way. The difference is this. A website has to turn a visitor into a buyer once; conversion is the number that matters. Nearly everyone using your app has bought from you before, and most are already loyalty members. The app does not have to talk them into trying you. It has to bring them back sooner, more often and through your channel rather than someone else's.
So a mobile app makes its money in 2 ways, and you can put a figure on both.
How often people order: a member who orders 11 times a year instead of 8 is worth about a third more to you, every year, and you did not pay to acquire them.
Where the order comes from: delivery aggregators take an effective 30 per cent and more per order once ads and promotions stack up, based on analysis of Australian marketplace fees (Oddle, 2026). The same order through your own app keeps that commission, and it keeps the data: who they are, what they order and when they stopped. One order in your own channel is worth several times the same order on an aggregator, because you keep the margin and you can bring that customer back without paying for them again.
If the mobile app will only bring in an extra $500,000, do not spend $400,000 building it. If it has to bring in an extra $20 million, $250,000 will not get you there. Spend too much and the return is not there. Spend too little and the mobile app cannot do the job.
Run your own numbers through those 2 things over 3 years, at your margin. That is what the mobile app has to bring in. That figure, and nothing else, should set your budget.
9 questions that decide the investment
Two QSR groups the same size can land at opposite ends of the range. It comes down to these 9 questions. Work through them before you go to market, not after the quotes land.
Which POS do your stores run, and does it stay in charge of pricing and availability? Nearly every QSR app is built on top of the POS already in the stores. How the app talks to it, and who holds the master copy of the menu, is one of the biggest costs in the project.
What happens to the loyalty program you already have? The question is whether every member and every point balance comes across without anyone losing anything, and whether the program works harder in the new app than the old one. Moving a live loyalty program is careful work and it is rarely priced properly.
How do customers order from you? Collection, delivery, order at table, drive-through, kiosk, scheduled orders, catering. Each is a separate flow with its own build and its own testing.
Are the aggregators inside your app, or outside it? Letting a customer order delivery inside your own app, and getting that order to the right driver, is a different job from sending them out to Uber Eats.
How do people pay? Cards and wallets in the app is one thing. Paying at the table, or on a terminal in store, brings hardware into the project.
Are the stores franchised, company-owned or both? A franchise network needs central control of menus, pricing and promotions, without every store needing its own set-up.
How far do you want to personalise the experience? This is where repeat orders come from: recommendations built on what a customer actually orders, one-tap reorder, rewards surfaced at the moment they would be used. AI has made far more of this possible than 2 years ago, and it is the biggest opportunity in most QSR apps today. It also has to be built into the foundations. It cannot be added later.
How much control does your marketing team need? Weekly offers, home screen banners, promo codes, push, photography, dietary flags. The more your team can change on their own, the faster the app improves; the less, the more every campaign becomes a developer ticket and a 2-week wait. Where that control lives, and how deeply your DXP, CRM or CDP connects into the app, is a real cost and it is usually left out of the brief.
What can you measure right now? If you cannot see where customers drop out of your current app, you cannot prove what a new one is worth, and you will struggle to get the money for the one after it.
The 3 levels of investment
Over 15 years of building QSR mobile apps, these are the numbers we see. None of the 3 is the wrong answer. Each is right for a different kind of business. You waste money when you buy the wrong one for the job your app has to do.
$250,000 to $450,000
The cost of a revenue-generating ordering channel for one brand
A mobile app and web ordering for one brand, built together so they behave the same way, running on the POS you already have. On the payback test, that is a channel that can earn $750,000 to around $2 million over 3 years.
Most of the money goes into the things that decide whether someone orders, and whether they come back.
Signing up: fewer fields. Apple and Google sign-in. A verification code that fills itself in. No asking for a date of birth you are not going to use
Ordering: last order reordered in 3 taps, not 4 screens. Real photography, clear options, dietary flags. Upsells that offer food, not only drinks, at the point where people say yes. A cart that keeps the order when you change store or switch from delivery to collection. A delivery address flow that takes 3 steps instead of 8. A checkout with nothing on it except the order and the payment
Loyalty: points balance and next reward visible every time the app opens. Offers added to the cart from the offer itself. Rewards that work at checkout without an error the customer cannot fix
Under all of it sits the connection to your POS. It is a large part of the cost and you never see it. It keeps the menu right in every store, puts the order in front of the right kitchen, puts the points on the right account and squares up the payment.
At this level the app earns by taking orders back from the aggregators and giving the members you already have a reason to use it. For a lot of QSR groups that is the whole business case, and it is a good one.
What it will not do is show different things to different customers, or let your team test what works. It will serve the demand you have. It will not grow it. That is what the next level is for, and why it costs more.
$450,000 to $700,000
The cost of an app built to grow the channel, not only serve it
Everything above, plus the parts that make people order more often and spend more when they do. The earning test here is $1.5 million to $3.5 million over 3 years.
Personalisation: a lapsed member, a weekly regular and a first-timer see different things when they open the app. Recommendations come from what a customer has ordered before and what is in the cart now. This is the biggest lever on repeat orders that most QSR groups are not pulling
Testing: your team runs 2 versions of a home screen, an offer or a checkout, and keeps the one that sells more. Without this, every decision about the app is somebody's opinion
Offers your team can run: a home screen space marketing changes weekly on their own. Offers that apply themselves to the right cart, instead of relying on a code the customer has to remember and type
Abandoned carts: most people who fill a cart do not check out. If you can see who they are and reach them while they are still hungry, some come back. For most QSR groups this is the biggest pool of money they have never counted
Loyalty that changes behaviour: not only points in and points out. Challenges, streaks and games that give a member a reason to order this week instead of next month
If you run more than one brand, this is also the level where a second brand runs on the same platform rather than needing its own.
None of this is a feature you can add later. It is built into the foundations or it is not there. An app that was not built for personalisation and testing will not take them afterwards, however reasonable the request sounds when you finally make it.
What it will not do is handle the in-store side of the business. Kiosks, order at table, drive-through and catering are each a real piece of engineering. They are what the level above is for.
$700,000 to $1 million and beyond
The cost when the app is how the business trades
Digital is your main channel, not a slice of it. The earning test is several million over 3 years, and groups at this level usually already know their number.
The app handles every way a customer orders: collection, delivery, order at table, drive-through, kiosk, scheduled orders, catering. Each is its own flow, its own connection to the POS and its own list of things that go wrong on a Friday night. Behind it is a large store network, often franchised, with its own systems - and the data has to be current. At this size, a team that can see what happened this morning and fix it this afternoon is worth more than a team that gets a better report next month.
Customer experience is the biggest driver of return at this level, not the smallest. When most of your revenue arrives through the app, the difference between a good ordering journey and an average one is worth millions.
What this is not is a way of showing ambition. If digital is 8 per cent of your sales and the aggregators carry most of that, spending here is the same mistake as underspending, made in the other direction.
Why the numbers you are hearing do not agree
When 2 agencies quote the same brief at $300,000 and $750,000, the obvious reading is that one is sharp and one is expensive. That reading is usually wrong. They have priced 2 different things, and the brief did not say enough to stop them.
One has quoted the mobile app. Screens, ordering, a payment gateway, a loyalty tab. The work is real and the result can be very good, but the shape of it is known in advance, so the number is confident and low. The other has quoted an ordering platform, and the app is the part of it you can see. Underneath is the layer that talks to the POS, keeps the menu right in every store, puts the order in the kitchen, puts the points on the right account and squares up the payment. That layer is most of the engineering and none of the demonstration. It is why the cheaper number is cheaper, and why nobody notices it is missing until the second year.
Both agencies have answered the brief. They have answered different questions inside it.
What AI has changed, and what it has not
Most groups going to market this year assume AI should have made an app cost less. It has changed some of the work and left the rest alone.
What it has changed is how quickly a design gets in front of a customer. Working prototypes are built in days rather than weeks, several rather than one, and real ordering journeys are tested with real customers before the money is committed. Ideas get proven or dropped while changing them is cheap, and testing now finds the awkward cases before customers do.
What AI cannot do is decide what the app should be - what your customers want, what it needs to earn, which of the 9 questions matter most to you. And it cannot design software that keeps working when every store is busy at once and a system it relies on stops responding. Left alone, AI produces something that works in a demo and fails in a real store. The architecture, the engineering judgement and the accountability after launch are still human, and on an app carrying ordering, loyalty and payments they are most of the bill.
What happens when you spend too little
A group that spends $250,000 on an app that needed $600,000 does not find out at launch. The app ships. It looks good. The rating recovers. Orders come through it, and for the first year the numbers move, because almost anything is better than what was there before.
Then the numbers stop moving. Digital share settles at a new level and sits there.
So the team tries the obvious things. Show a lapsed member something different from a weekly regular. Run 2 versions of the home screen and keep the one that sells more. Chase last night's abandoned carts while the customer is still hungry.
None of it is possible. None of it was built, and none of it can be added to an app that was not designed to carry it. That decision was made 2 years earlier, in a meeting about price, and nobody in the room knew they were making it.
Now there are 2 choices and neither is good. Keep extending an app that cannot take the weight, and pay for that every year. Or build it again. Most groups build it again, which means paying for 2 apps and getting the use of one, and losing 18 months in between. The 18 months usually costs more than the second app.
3 ways to genuinely bring the number down
These 3 save money without leaving you with an app that cannot do the job. Two of them cost an agency work, which is why you do not hear them often.
Check whether you need a rebuild at all: a lot of what gets blamed on the app is not the app. A sign-up form asking for things you do not need. A minimum order value nobody has questioned in 3 years. A delivery address flow that loses people at step 6. Your best offers sitting on the website when they should be in the app. Delivery priced no better in your own channel than on Uber Eats, which gives the customer no reason to switch. All of that can be fixed on the app you already own, for a small share of what a rebuild costs, and it will move your owned channel share this quarter - and give you the numbers to prove the rebuild was worth doing. Sometimes the honest answer is that the rebuild is next year's job. That is the cheapest outcome available and it is worth looking for
Build the app and the web ordering together: web is usually the bigger share of your owned channel revenue, and it usually gets treated as an afterthought. Built together, they are one build, one round of testing and one roadmap, and a change goes live on both at once. Built separately, you pay twice and then keep paying, because every change is made twice and the two drift apart until customers notice
Only build the ways of ordering you actually use: kiosk, drive-through, order at table and catering are each real engineering, and groups routinely put all of them in the brief because they are on a roadmap somewhere. Build the ones carrying your orders today. Add the others when the volume is there, on a platform designed to take them. That is a real saving. Cutting the foundations that let you add them later is not
The question to take into the room
When the quotes come back far apart, the instinct is to ask the expensive agency to justify itself. Fair, and wrong: the answer will be a list of features, and it will sound like a sales pitch whether it is one or not.
Ask every agency on the list this instead. Which parts of this quote are foundations, and which parts are features? And which of the two can we add after launch?
A good answer draws the line cleanly. It names what sits in the foundations - the POS layer, the data model, the personalisation plumbing - and tells you which features that foundation will carry later, and which it never will. A bad answer says everything is modular and can be added any time. The answers will not be similar, and that is the information the price gap was hiding.
If you are working out what your own mobile app should cost, tell us what it needs to do and what it needs to earn, and we will tell you what we think it should cost and why. Talk to us about your project.
Learn from us
Join thousands of other Product Design experts who depend on Adrenalin for insights
