My 2 months as the ‘Safaricom Head of Innovations’

Feb 12, 2013 25 Comments by

It was all like a dream

310pm. It was a warm Thursday afternoon that I decided to shut down Faustina (My Samsung Ultrabook). I had been coding for over 4 hours straight. The much anticipated AngryKenyans game was coming up OK. But the perfectionist in me kept on delaying any tweets or releases about the game.

I did not like the UX in the game from my Lab. And Sue, my favorite designer, was still to send me the latest splash-screen image. So, my fans were kept waiting. But my mind was tired. I needed to run to Qz for a training session, as a member of the National Pool League. I did a last-time sync on my S2 using the Nailab Wi-fi, Dropboxed all my data and then rushed to my parking spot at the basement where my sweet Henrietta was parked. My phone buzzed. New Email from Safaricom HR.

Hmmmn. What have I done this time? I read, carefully.

“Dear Mr Idd Salim,

Welcome to Safaricom.

We are pleased to announce that after a rigorous month of interviews and profile checks, you are our choice for the Head of Innovations at Safaricom. You will have a 3 months probation in this new role starting Monday.

Kindly report to the HQ at 8am to finalize the working contract.”

I could not move for 5 minutes. OMG!! This was really happening. Finally. This was good news. No. Good news is being told that your elec bill is 200 bob. This was HUGE news. Everything I had preached about on my “what Safaricom should do” was now flat on my desk. “Do it yourself, now that you got the power.”, the Kenyan tech scene sang to me in unison. In my mind.

I had a polite weekend. Got about making sure I had “official clothes” and shoes. Those jeans I used to wear to iHub (when I used to go there) and Nailab, would now be tucked away in a box. Only to see the sun over the weekends.

Monday : 9am
Reached the HQ at Waiyaki. Was welcomed by the Head of HR and shown to my desk on the 4th Floor. After 2 hours of orientation and briefing, I finally settled. This is my first day. I needed to hit the ground running.

Setting the base:
So, I went through my long list of the things I would have wished to do. I finally had the chance. I chose 3 of the “10 things I would happily die after doing”.

1 – Aggregated URL-Based Data Revenue Sharing
Safaricom have always been adamant to share Data revenue with developers. This is mainly blamed on bundles and differing rates of data due to different data plans. Ofcourse, this has directly led to Kenyan developers not being as innovative as they could be. BuniTV and the like were making GBs of local content per day. But still never got a dime. I decided to sort this out.

My solution was simple. There was need for Safaricom to maximize revenue on data. There was also the need for the developers to get paid for their data apps. I created a CDN on the Safaricom cloud and called it DataMint. DataMint’s purpose was to aggregate all mobile data streams from all developers, to one SINGLE entry point for easy accounting.

So, every developer just needed a unique APP ID which is used to fingerprint their data streams.

For example Developer ABC, with the fingerprint AABCCDDEFF, who wants to serve a video content “abcd.3gp” will just need to pass this to DataMint e.g. datamint.safaricom.co.ke/AABCCDDEFF/abcd.3gp. Simple. Then, each developer had a web interface from which they could see their earnings in real-time. All payments were made on-demand and were payable via Mpesa and EFT.

All the data rates on the DataMint were discounted, to encourage more data consumption.

In terms of billing, All data that passes through the DataMint will be accounted for in the following model:

  1. All data charged outside bundles – at KSHS 8 per MB – Revenue Shared 45:55 at the favor of Safaricom.
  2. All data charged on the 10MB for 8 Bob daily Bundle – The 8 shillings Shared 45:55 at the favor of Safaricom for each 10 MB. If the time expires before the 10MB is done, since the money is already charged, the revenue shares remain as above.
  3. The other data bundles follow the model above

2 – Open USSD Platform

The amount of money made by SMS campaigns is astonishing. This is from a population largely describes as “illiterate”. Not suprising, a big percentage of SMS revenue is made “accidentally”. Someone requesting that song by Kidum K1234, most of the time will get that song by Kiyapi K11234. This is due to user error. And errors are always billed.

The industry would rather bill you alot ONCE on error, and kill your confidence in the system, rather than bill you accurately in little bits for a long time. Now, opening USSD to the public the same way SMS is open, would have untold applications.

Accurate data requests even by the most illiterate people. Useful apps. Opinion polls from more that the 1000 Kenyans that IpsoSynnovate can reach. Endless

3 – 13Apps App Store

“Salim, how do our Apps get noticed?”, came an email from a coder. I adopted a walled-garden model. We have over 900k Android devices in Kenya and over 6M Java Data devices. My aim was to support and guide local developers on the PROPER development of SCALABLE apps.

Not 100 apps. Not 1000 apps. Just 13 apps. Just 13 quality apps

The aim was to create such a setup within Safaricom that, No Kenyan apps were put on GooglePlay. No Kenyan Apps were put on the OVI Store. The 13 Apps were hosted on DataMint AppStore

The Apps needed to have the following featuresets:

  1. Can attract usage from at least 30% of the Android and Java phones. – USP
  2. Can attract usage from a user at least 3 times a week. – Addictive
  3. Can generate residual data/SMS traffic per use. – Monetizable

Inviting the techies:
So, every 2 weeks, we held hackathons. Alternating between iHub, Nailab and 88MPH. My mandate was simple. Oversee the development of the BEST apps in terms of design, UX, network usage and revenue inclusion. As a seasoned coder and server admin, I brought alot to the table.

Ofcourse, I personally did QA and code-checks, made sure the code was writen well, did DB checks and optimization with the coders to make sure the app ran as well with 1,000,000 people, like it would with 100 people.

Apps were created and they conformed to all the users wanted. A light, nice-looking, never-crashing, functional app with an inbuilt user-stickability incentive and a solid and residual revenue model.

6 weeks later
In my office now. Moved from the desk. Chilling coz i’m thrilling. Sipping something cool while thinking about pool. Looking at the 48 Inch Samsung TV screen that notified me on serious and critical KPI metrics like user fatigue, revenue drops and surges to help me strategize.

The data numbers started rising. The developers were all now singing “Naafurahiaaa…. Niko na Mfuko imefura”. Kenya was now on the tech-map for Africa. As what it could be. We no longer had google-maps clients as our “biggest talk-points”. We had REAL systems. A Kenyan Skype, over a vLan. SuperComputer powered sentiment analysis tools.

All bliss. Just before Musyoka, my errands attendant, came to my office and announced that we were having a fire drill. Loudly, the siren rang. Just 2 minutes before I came back to my senses. In my bed. Then realized it was no Siren. I was rudely awakened by my S3 Alarm. My God! It was only a DREAM!

A good dream of what could have been. A bad reality of what isn’t. Rushed to the shower and prepared myself for another day of self-enlightenment and self-challenging in code and data models at the Lab.

Back to code.

Wazi.

Personal, Startups

About the author

Coder, hacker, inventor, pool guru.

Switch to our mobile site