Things you might not think about when planning your app

As we enter into October 2025, I'm reminded that winter is a busy time for in the apps world. The kids are back at school so a lot of businesses are finalising their app plans and reaching out to app developers for advice and quotes, so they can make their business case.

Launching a successful consumer app requires careful consideration of factors that many teams overlook. This guide covers eight critical areas, from app store ratings and analytics to security and technology choices. These can make or break your mobile strategy. Based on insights from managing apps for 500,000+ users, we outline the key questions every CTO and Product Director should ask before development begins.

Don't sleep on your star rating

If you're like most, you'll want your app to have a good star rating. According to research, 4.0-4.9 is the sweet spot. Not everyone realises that you need to put a bit of work into achieving a good star rating. You can see this challenge in any category of app. For example, I studied eighty Smart Home apps for controlling household lighting and appliances, and 48% of them failed to achieve 4.0 stars or above. 

Here are a few examples.

These aren't small, cash-strapped brands either. Needless to say, low-scoring apps have some negative consequences. 

• Fewer downloads: Apps that are not frequently rated and reviewed won't be prioritised in app-store searches, meaning people browsing and searching the app store won't find you. You'll miss out on free downloads.
• Less revenue: A poor app store rating means the people looking at your app store page are less likely to download, leading to fewer downloads and revenue.

So why is this so hard, even for big companies? Well, there are three parts to it:

1. People rate the service, not the app: This means that if you're a taxi company and you're constantly late, an amazing app will still get poor ratings. Or if you're a manufacturer and your smart device keeps disconnecting, people will leave bad reviews.
2. Quality counts: A clunky in-app experience will still generate negative reviews. Achieving a good quality interface doesn't happen automatically. You need to invest in it. For example, you can run user tests or ensure you're working with senior designers.

Back in 2023, we developed an app that people raved about during user interviews, but our rating was only 3.2 stars. We realised that the haters were more vocal than the superfans. So, we created a tool called Appero to encourage the superfans to leave a review. In the space of 6 weeks, we accumulated thousands of new ratings and reviews, and propelled our star rating from a mediocre 3.2 stars to 4.3 stars.

A high star rating is more costly and time-consuming to obtain, but it will increase downloads and improve brand reputation. So ask yourself: can you get away with three stars, or does it have to be four stars or more?

Use product analytics to improve app performance

Modern app development teams working with companies like Uber, Nike and Asda put a lot of effort in to measuring customer behaviours inside the app. They can literally go into their analytics platform and see where customers abandon their purchase, or where they spend too much time on tasks that should be nice and quick. 

Check out the example below where we're measuring how many people get through the first screens and purchase a subscription for a health and nutrition app.

Using this as an example, ask yourself: what customer behaviours do I need to measure? What's the make-or-break user journey that must be slick in order for my app to be successful?

For example, is it critical for people not to drop out during registration? Or if you're in the wellbeing space, perhaps it's how many people log their first journal entry or set up their goals?

With analytics in place, you'll find your app development team can improve an app because they know what's happening under the bonnet. 

In terms of improving workflow, once you have these analytics, you can start to use them to guide your development plans. Too many people abandoning at the registration screen? Then perhaps you should invest in re-working that, or using psychology to create more desire or intent.

Now, I know you're probably busy for all this and don't want to be in the weeds with all these details. So it's about making sure your product team have the right tools and processes to drive success through product analytics.

Extra things to know about your customers

If you're a product director or CTO, you probably fully understand your industry and audience. However, you may need to dig a little deeper when it comes to designing an app or digital platform. Here are a few useful things to know when designing mobile apps.  

• What phones do they own?
• What are their accessibility needs?
• Where are they, geographically?
• How often do they need to use your app?

For example, some apps get 70% of their revenue from people on tiny, cheap phones that they only upgrade every two years. So you have to design for them. Or, some older people living with diabetes are very likely to have visual impairment, so you need to meet WCAG accessibility standards. 

What technology choices do you need to understand?

If you don't work with technology teams often, it's worth having a top-line understanding of what's going on with your app. After all, apps are always built on a "tech stack" - meaning the technology choices behind the app. And those choices influence a few important things:

• Mobile tech stack influences development cost
  
  E.g: React Native or Flutter is cheaper than building native Swift and iOS apps. Relying on 3rd party providers such as RevenueCat for billing will reduce initial costs (but add more later). 
• 3rd party dependencies influence operating costs
  
  Most 3rd party dependencies such as OpenAI, OneSignal push messaging, Azure hosting or Amplitude analytics will incur a potential running cost to your app. This can quickly add up.
• Tech choices affect ease of future hiring 
  
  Sticking to mature and well-established technologies is a good starting point if you want new developers to join your team, or a new team to replace your existing one.
• Not all tech is scalable
  
  You don't need to know all the detail, but simply asking the question "How do we cope if we get 6x more users?" is a good thing to know. We once had to rewrite a middleware solution to cope with massive scaling. This isn't always a bad thing, but it's good to know so you can be prepared.
• Customer experience quality can be reduced by technology choices
  
  Technology choices affect how good your experience can be. Native apps are typically the gold standard but most expensive to develop. Flutter is pretty close in our view, but there are compromises in user experience. For example, Android and iOS users will see the same style of app. Most won't care, but it's good to know your limits. 
  
  
• AI accelerates development, but also presents IP risks
  
  Since writing this article in 2022, a few things have changed in the past 3 years. The biggest change is that many teams are adopting AI-accelerated development. And that brings in new risks. Your top-secret algorithms and IP can literally be shared in to a public LLM for all to see if you're not careful. So check their AI policy and make sure you're comfortable with how they're using it (here's ours).

How to leverage AI in your app

In 2025, AI is top of the CEO agenda. Not just using it for productivity, but to enhance customer journeys. My personal belief is that you should first shoot for some low hanging fruit to get quick ROI from AI.

As food for thought, we recently did this with Carbs & Cals in three different places.

1. Using LLMs to scan food labels, to reduce data entry time for customers
2. Using LLMs to scan food photos, to help customers log entire meals in seconds
3. Using LLMs for data quality control, to help the back office data team quality check and improve data faster

Carbs & Cals isn't a trivial app, it's endorsed by NHS professionals and helps almost 500,000 people live a better life with diabetes by giving them reliable data (data is from our app intelligence tool - Apportunity). 

For what it's worth, you should dip your toe into AI by looking for quick wins that bring you some fast ROI. And also expect to invest a little in experimentation - it took us a while to find the right LLMs and SLMs that produced fast response times and accurate results at an operational cost that was acceptable. 

Try to understand what devices your audience use

At the time of writing, iOS and Android are equally dominant. If you're developing a consumer-facing app, you will likely need to launch on both (unless you're looking to develop a quick Proof of Concept). 

Then you need to think about device popularity. What screen sizes and devices will you design for? There is a vast array of popular phones in the UK, and it's useful to pay attention to the devices your target audience are on. 

As an example, we develop and manage an app used by older people, many only buy a phone to use this one app. So they don't spend a fortune. This means that we have to test the app on fairly cheap phones that might have small screens and run much slower than the flagship phones. It's important that we keep these customers happy, so we have to optimise for them. 

Image from our Badly Drawn Mobile guide.

The downside of this is that it increases our quality assurance costs, because we're testing and optimising for more types of device.

Another thing to remember is that older and cheaper devices won't be able to support on-device AI use cases. They're not capable of running a local LLM, which are RAM and power hungry. If you plan to use AI and LLMS to create convenience and value for your customers, then you'll need to use off-device LLMs.

Over 1,600 business are hit by cyber crime every day

In the UK in 2025, Gov.uk reports that 612,000 businesses have reported being a victim of cyber crime. That's a whopping 43%, and over 1,600 businesses getting attacked every day on average.

If you're a medium-to-large sized business, there's a 64%-74% chance you'll fall foul of an attack. 

Your app forms part of your attack surface. NowSecure states that 85% of mobile apps they analysed have vulnerabilities that could compromise security or personal data. So, it pays to have a plan here. Even moreso if your app handles payments or sensitive customer data. So when building an app you need to consider: 

• Should we budget for a PEN test?
• Should our app developers to work to security standards such as OWASP?
• Should we consider continuous security sweeps with tools like NowSecure

Most businesses aren't doing enough about cyber security, so I'd advise it comes up as a consideration in your mobile app or AI app planning.

Roadmap

What features or bells and whistles can you cut? Developing an app can take more time than you’d hope, so it’s helpful to know what you can aggressively cut for your first release, just in case.

The good news is that most of these questions can be answered in minutes rather than hours or days.

Some are super important and aren't reversible, so give these questions enough thought to ensure you set off in the right direction. Post in the comments below, and sign up for the Pocketworks newsletter for regular mobile tips.