I’ve come across some government agencies and other businesses claiming they’d spent millions of dollars developing an app. How can developing an app could cost so much money?
In: 496
When you get down to it, there are a *ton* of ways even the simplest app can be written. For a small or simple app, it’s often not too hard to track down a developer who can do the work.
Writing the app is often one of the *simplest* parts of the process, though. That app actually has to do something. Since this hypothetical app is a little more featureful, you’ll need a more capable developer. You’ll also need to put the app together with the hardware. Which hardware the app is running on might matter: do we want to rely on a server? If so, there’s another developer who’s got to write that server app. What about client devices with limited memory and power?
Those client devices probably are running different operating systems. You need to have a version of the app (and maybe the server software, too) for each OS. Are there different devices with different OS versions? Good luck!
Who will provide the cloud service? Is it the same agency that’s providing the server app, or are you sending your app’s data to Amazon, Google, or Microsoft? Maybe you’re hosting an entire data center on a whole bunch of machines and networking gear you’ve purchased yourself?
Oh, and one more thing: the app *needs to work*. More specifically, it needs to work well once you’ve got everything deployed. This means writing a functional specification, and actual test plans that will prove that the spec is met. It’s helpful to have a manager who specializes in this (some people call these project managers) to coordinate the work that everybody’s doing and keep everybody on the same page.
Finally, you’ll need people to provide support. Not just technical support, but helpdesk-style support. People to answer phone calls and e-mails as well as the myriad social media interactions you’ll (hopefully) be getting.
All of this costs money. The more of it there is, the more ways there are to screw things up, add complexity, and make the project grow beyond what could ever be done by one person. The more critical the app is to the business, the more people you can expect to be involved and the more expensive the work will be.
Let’s say you have a team with 5 developers, one designer, one product manager, one people manager.
Each of them cost $200k/year, fully burdened.
That is a total staff cost of $1.6M per year.
That is just one team. A heavily featured app might have dozens of teams.
As well as what other people have said, the back end infrastructure to support an app can be very expensive to both develop and run. Suppose google didn’t exist and you wanted to make an app to find web pages relevant to a search. The app itself could be quite simple, but think about what’s involved in crawling and indexing the whole internet so that search results are good.
The UK govt claimed they spent ~£38 BILLION on a tracking app during covid. Corruption is where the money went
It’s surprisingly easy to spend that much money:
Security is probably the biggest: for a government or other major organisation the security has to be rock-solid, and that means all the right design, development and testing to cover things like account management/resets, multi-factor authentication, on-going penetration testing etc. – you cannot compromise on this and Security experts are never cheap.
Data protection: most organisations want to use the app to gather customer data to improve their marketing and customer profiling – what data can you gather/store, for how long can you keep it, what do your T&Cs need to say on sign-up? Can customers ask to be forgotten, can they see what you hold on them? For what will you use their data, because you will need to make that clear up front.
Providing something actually enterprise-strength also requires a shitload more designing than just developing the app. How will you ensure that it works 24/7? You’re looking at a back-end infrastructure that supports the app with proper failover procedures, more testing, hot standby environments etc. Add in other non-functional aspects like performance, latency, scalability etc. – all need to be considered.
Most apps are not just self-contained apps, they are all reading/writing to/from back end services all the time. Which new or existing APIs will the app depend upon? If you develop the app but your customer details API is slow and falls over above 100 calls per second then you’ve got work to do there. Most organisations still rely on slow-moving batch data and often need to upgrade their data provision to something much more fast-moving to support this kind of Digital interaction.
Big organisations will often spend months just in the user experience design – customer profiling, process flows, use case definition, market analysis, beta testing, design consultancies etc. Will it work for blind/partially-sighted users, will it work in other countries, can you change the language, will you have one app for your whole organisation or multiple? Do you have different brands under the same umbrella organisation? Do user accounts transpose to your website as well or are they separate for the app? What is the logical data model e.g. can one email address have more than one account? Can one address have more than one account?
Operational support: what will users do when the app doesn’t do what they want, or they get locked out of an account? You will ultimately need a team of human beings who can deal with problems that can’t be fixed within the app. How will bugs get fixed? How often will you release a new version of the app with bug fixes or improvements, and will you force the app to udpate on the user’s device each time?
Operating model: lots of organisations launch an app but neglect to think about what this does to their business. If you own a bar/restaurant and launch an app, how will the kitchen staff receive the orders? How will bar staff know to check a screen for app orders coming in, and where to those orders sit in the queue versus people who ordered at the bar? How will the restaurant let app users know if they run out of a particular meal? Will you need table numbers on all your tables now? How will app users pay the bill? Will the bar staff still get tips if everyone pays on an app, and if not will their pay need to change? All of this has to be worked through, and potentially rolled out in the form of training and operating changes for your staff.
That’s off the top of my head in 5 minutes. I reckon there’s loads more than this.
Understanding the cost and complexity of creating an app compared to the sheer number of redundant and unnecessary apps in existence is ironic.
People/users are curious and unpredictable. Poorly written code is even worse. Combine the two and your app is not going to work.
Fixing the latter is a matter of several hundred thousands dollars of professional software engineering.
If your software can handle the former, then you stand to make a fortune. However, that is very hard and thus very expensive.
Besides the cost of the app, the guvment can be really inneffecient, paying multiple people to do or oversee the same job. One agency I had to see a nurse hired by the guvment from a third party just to qualify me to see a nurse at the agency.
You have government business owners that run a project from their side, probably 2 or 3 staff. You have other department stakeholders that provide input, and they’re not doing that for free. You have an IT program manager. You have an app dev. You have a system analyst. You have management within IT. Devs for all the internal systems that need to talk to each other (and these systems are often old within government, so those devs that can code those languages can charge more) user experience design staff, you have communications and media who aren’t doing that for free. The list goes on. What seems simple is actually quite convoluted.
They will hire an external consultancy. The consultancy will want to do a discovery period to break down exactly what the app needs to do, how it will benefit customers, how each feature works, what it looks like and the priority of which to build and release first.
Then there’s the technical side, what does this interact with to do stuff, are there apis available, do we need to build an integration and or orchestration layer, how do we handle user authentication? We need architects, developers and specialists from the systems were integrating with.
What and how do we deploy the systems that we build? Is there a governance process for creating new test environments? Do they have test versions of all the prod systems and how well do they mirror prod? What about test data? Do they have a UAT or infosec security process we need to follow?
I could go on, but you can see how this portion alone can have dozen or dozens of folk from both organisations and can easily take months.
Add in the delays to get the right people together, uncovering hurdles and here we’ve had 50 odd folk working for 6 odd months and haven’t even started building yet. Add all the time for procurement to get finances flowing lol.
Source: CTO for a consultancy that builds apps for large organisations.
Latest Answers