How we built a world class app under $500.00.. for iOS and Android
In 2013 my colleague Ghartey and I were having a casual conversation with the then co-founder of Saya Mobile, Badu about a simple app for meeting interesting people around you. However, this was just a onversation because Badu was busy finishing up development of the Saya iOS app back then.
Fast forward 2014, Saya Mobile was acquired by Kirusa, Badu didn’t join Kirusa and we decided to do something fun since we really enjoyed working with the Saya team as their designers. We didn’t have investors. We decided to bootstrap. This basically means cash in bank was US$0.0 from get go. One Design Guru + One Genius Programmer + One Content Guy was all the talent we thought we needed to pull this off.
What we decided to build was simple and somewhat based on our conversation the year before; a sort of a social app, that uses phone numbers to sign up users and help them meet new people through their contact social graph. This sounded simple enough or so we thought.
We started prototyping and created a solid iOS prototype using Flinto to test our hypothesis. Friends and people we tested with thought it was a fully functional app and glowed with excitement. We were excited too and didn’t mind paying US$20.00/month.
Total Spend = US$20.00
What next? Branding, website and social media pages. Ghartey executed the entire branding from logo to website and native app designs in adobe illustrator and photoshop. We “borrowed” a few images from free stock sites and photos of friends.
We had already found webflow by then and created the first Beep website using their awesome tools on a $35.00/month subscription. Mean while, Badu had already started working on the iOS app.
Paid Services = Webflow/Month + Flinto/Month
Total Spend = US$ 2o+US$ 35 = US$ 55.00
We realized two hurdles that could do us in. First, since we were eliminating email and social logins in favour of phone number at sign up, we either had to pay for SMS authentication to be sent to users to sign up or push that cost to users to text us a code generated by the app upon sign up to authenticate them. The latter was the most undesirable option but it would also hurt our empty pockets if we were to go with the former. We had a few meetings with Infobip to explore the cheapest and most viable solution. All the solutions we got from inforbip and later Twilio suggested we had to swallow the cost of SMS.
The second hurdle; because we were using contacts graph, a user would only see her contacts who are on the app and then discover friends’ friends that she could connect with. What this means for our iOS app was that, the user’s (on an iOS device) contacts didn’t only have to be on an iOS device but they needed to have the app as well and so did their friends. This further complicated the problem of user base that apps within our circles faced. Android phones are also more popular among our target group in Ghana and Nigeria where we planned to launch our beta. We would have been better off launching initially with only an android app.
But hey, our iOS app was already looking great and so why stop now? We didn’t think twice about paying US$ 99.00 for the iOS developer programme either.
Paid Services = Webflow/Month + Flinto/Month + iOS Dev./Year
Total Spend = US$ 55.00 + US$ 99.00 = US$ 154.00
We integrated Mixpanel for testing our iOS app on free data points. We love all the analytics, features and insights we could pull out of Mixpanel but were not ready to pay just yet. When we started running out of data points, we slapped their badge on our website for more points.
Our first attempt at getting android was to contract a developer. We didn’t have enough funds so the idea was for Kweku to do design for the developer in exchange. That didn’t work. Getting an android app was such a bother for us that we decided to save US$5000.00 to outsource it. Somewhere along the line, Badu figured a way to build a native android Beep app. The stars were almost aligned. Kweku didn’t take any prisoners on the Beep android designs. After a few iterations our V1 Beep android UI was complete. We skipped prototyping and went straignt to production. As soon as we had a working android prototype, we integrated with Mixpanel for some more testing.
We have been using Crashlytics for awhile and are quite familiar with some of the twitter developer tools. I got a mail from Badu about a new tool kit twitter was providing for developers called Digits. If what I understood from their website was right, it meant app developers didn’t have to worry in the slightest amount about phone number sign ins, Digits was here to deal with the complexity of OAuth and SMS cost. All the heavy lifting in building our own login infrastucture and cost associated with SMS authentication was going to be provided for free. We just integrate Digits into our app and all users have to do is simply enter their phone number, and Digits will send a confirmation code to verify the number. To top it all, Digit had a friend finder feature that we can implement to help users find other Digits users in our app that they know through their address book. This would further guarantee the safety of our app but you know what they say when something is too good to be true.
I later found the Twitter flight keynote video sometime after and sent it to Badu. It would appear christmas was coming earlier than usual. Digits was for real!! For a better perspective, please read Badu’s blog about How we cut SMS costs using Twitter Digits which he posted right after we integrated Digits.
All this while our US$100.00 credit on Digital Ocean where we hosted our app was running out. Sorry, it actually ran out, and we were enjoying negative balance. We managed to top up with another US$100.00 to keep our servers running.
Paid Services = Webflow/Month + Flinto/Month + iOS Dev./Year + DigitalOcean
Total Spend = US$ 154.00 + US$ 200.00 = US$ 354.00
February — Ghana V1 Beep Release
We are happy to be here now and released our first version on both iOS and Android in Ghana to create a valuable community base of users and reach out to other parts of the world later.
This estimate is solely based on services we had to pay that were directly related to the development of Beep. It doesnt take account of all the man hours we spent iterating back and forth. We also treated Beep like a side project so we didn’t take any form of salary throughout the development period.
