As a busy founder, we easily miss some easy things that can save thousands of dollars in app development costs. My suggestions in this post give you easy access to a goldmine of crucial product insights. These insights are easy and you already have access to everything you need. You just needs some time and dedication to build a better product and save on app development costs.
I recently spoke to a pair of budding New York startup founders. They wanted to rebuild their local discovery app (deliberately not disclosing name). They launched almost a year back and are doing pretty good for a locals app (upwards of 100k downloads). For the past quarter, they started seeing lots of app quality issues (not necessarily due to scaling). Some of their friends with the technical knowhow suggested rebuilding the app to prevent further business breaking issues. Due to app quality issues, they were struggling with lots of negative user reviews impacting their growth.
We get into a lot of meetings like these where entrepreneurs have no other means than to yank the chains and do a complete rebuild because of app quality. This is a very hard decision. You can outsource app development and design or put your existing team on it. Either way, your app development costs go through the roof. You’ll probably have to halt any new efforts on the live product for a while. Just multiply all the problems and risks of managing one product x 2.
Make a habit of exporting and analyzing your app reviews every month for at least 3 previous app versions.
Our process starts with discovery and analysis. Lowest hanging fruits in such cases are app store reviews. In most cases, your users will tell you if your cart is broken or if the credit card processing is not working the way it is supposed to. If you are facing app quality issues, my suggestion is to go to the app store dashboard and get the list of all your app reviews. Do this for your current and for at least 3 previous app versions. Analyzing these reviews will help you pinpoint the problems.
Measuring the impact
Before jumping to the app store reviews, I wanted to see how much of an impact we were talking about? They were using Facebook and Admob (Google) to promote their app via incentivized app download ad campaigns. It was clear from their reach and impression figures that the targeting was fine and they seem to have figured out their audience. When I asked for the campaign details and to my surprise, they obliged right away. I guess, they also wanted to see where this was going. We got to run some numbers across 30+ incentivized campaigns in 2 networks for the last quarter. The result? More than $43 per install!
They knew the cost per app install was high but they didn’t expect it to be that much. Jack (name changed) later mentioned it was around $16 in his head. This was based on the glances he was making on Facebook and Admob dashboards. Interfaces do that to us sometimes. When things are so neatly and effortlessly presented to us, it is harder to get into a habit of digging deeper and finding out.
My suggestion to founders is to always keep an excel (/google sheet) with your campaign performance details. All major networks provide the ability to download campaign performance metrics in .xls format. It takes 5 minutes every week to collate the reports and you get the ability to easily run your own analysis for free!
Use tools like excel to analyze your long term ad campaign performance metrics. Its easy and gives you actionable insights for free!
Measuring the impact beforehand is important. It tells you the what to expect and how to measure it after the fix. Looking at the cost per install, i would happily suggest a rebuild of the app and not think about app development costs. Making your app stable is worth a lot more than paying $43 per install and watch uninstalls happen within days. There is one more missing piece – whats causing the users to leave?
Root cause and it’s solution
Next stop was the app store reviews, app store crash reports, and the app itself. This would really tell me the kind of quality issues they were having. Are the issues related to new OS updates? Did a new feature use a 3rd party plugin/library that is causing issues? Are the issues happening on devices from certain manufacturer/s? Spending the time to find the reason behind the app quality issues could mean the app doesn’t need a rebuild. This could save founders a great deal of money in app development costs, time and headache.
It was a pretty popular app. Getting through with all the reviews and crash reports did take some time but it was worth it! There was a goldmine of analytical data that quickly pointed me to the problems. I come up with a prioritised list of problems and we picked the top 5. These could be fixed easily and have the most impact on app development costs. In fact we could solve 40% user issues if we just fixed the topmost item – Issues with a 3rd party library and 80% of the issues if we did top 5. This gave the founders a straight path to solve the quality issues that didn’t require a rebuild.
Crash reports are not just for developers. As founders, you should analyze your crash reports regularly.
3rd party GPS assist library that was used by the developer of this app was incompatible in Samsung phones with new version/update of android operating system. This is a fairly common occurrence. Phone manufacturers use their own Android ROMs to enable/disable or fine tune features suited to their hardware. This also makes their software distinctly their own. This causes issues with non-approved 3rd party libraries that are not updated regularly.
There’s nothing wrong in using repos from Github. Why would you spend all the time to build a GPS assist library and increase app development costs when you can fork the code from someone else? You should! but do so carefully. If you are not careful, you could land i rough waters like these. Not even a month ago, we had a similar issue in another app where a 3rd party timer library has problems working with HTC phones.
Due to faulty crash reporting setup (re-implementing crash reporting was the second suggestion), the issues were not getting logged for the developer to fix. In many cases, developers do not view app store reviews regularly. In this case only the founders had access to the Google play console.
Founders largely ignored app store reviews like these. They thought that such issues were due to faulty user devices or GPS not being able to get a signal. This was found when we analyzed app store review report, Android crash reports, OS version and publish dates. I’ve seen a lot of issues due to sub optimal implementation of 3rd party plugins. It probably demands its own blog post as well.
Avenues for your users to speak
Apart from fixes to the issues, i had another suggestion. Implement avenues for users to speak and feel that they are being heard. When users face issues in the app, app store is the only place to make themselves heard. Putting an in-app customer service/helpdesk SDK is a simple step. It is a tremendously useful solution to analyze quality issues. You get a pulse of what your users think about the app. You’ll get a clear action list enough to guide your developers.
We like Helpshift. Its a paid software but its benefits far outreach the savings you will do in app development costs in return. It has a real time messaging system for your users to communicate freely with you. It is like talking to your friends in WhatsApp or Facebook messenger. There is a FAQ system as well. When users send their issues, you get useful metadata like operating system, app version etc. To help diagnose the issue quickly, you can even choose to pass system variables like user_id in the metadata. If users can have founder’s ear for their complaints, no issues will go unnoticed for long.
Key takeaways
- Keep a check on your 3rd party plugins
- Setup proper crash reporting in your app
- Measure your ad campaign performance for the long term
- Keep a record and analyze your app reviews for previous 3 versions
- Integrate in-app messaging so that your users can submit direct feedback
