The Truth About Safaricom's Per Minute Data Tariff

This morning I woke to a full page advert in the dailies announcing that Safaricom had launched a new data tariff for KSh.2 per minute billed per second. Instantly several questions popped up in my mind, many of which were gratifyingly answered by @kachwanya's and @mkaigwa's excellent posts.

The concept of per minute browsing readily intrigued me because of my rural cyber cafe exploits as well as my general interest in Safaricom's data products. Before I could devise a strategy on how these new tariffs would impact my internet hungry businesses, I decided to give it a test trial. I figured a 10 minute session (presumably worth KSh. 20) should be sufficient to answer questions about speed and reliability. Below are the results of my test trial, which I should warn are not scientific and it would do well for you to carry out your own trials.

1. Deciding that a 3G modem would get the job done best, I first made sure that my bandwidth balance was expired. This was done by sending a Balance check request. I quickly confirmed that the my data balance was zero and airtime balance was KSh. 0.26.

2. I then sambazad KSh. 20 worth of airtime to the modem. I was however outraged when KSh. 10 was quickly appropriated to an Unlimited SMS service which I have never subscribed to. This was a fight for later for now I was focussed on my test, so I sambazad another KSh. 10 airtime to get my desired balance of KSh.20.

3. I sent the message ON to 142 and a minute later got a message that I had been activated on the Browse @2/- data tariff.

4. To test the time, I decided to use my stop-watch and started it immediately after Connecting the modem to the internet. (I should indicate that the modem connects to the internet approximately 5 seconds after clicking on the connect button).

5. Now came the good part, how much could I do in those ten minutes. I decided I was going to throw everything at it plus the kitchen sink:

• I had 10 browser tabs open, some with minute by minute refreshes.
• I started a 700 MB torrent download
• I opened a 4:42 minute YouTube video
• I attached to send a 5MB email through Outlook and
• I started a Windows Update of 16MB

6. I was running this test in Nairobi CBD where my modem gets full bars in HSPA mode. Technically I should be able to access 7.2 MBps speeds but I usually max out at about 1MBps. I happily noted from the modem's statistics page that I was hitting the 1 MBps. Based on that only I can say that this tariff doesn't have any bandwidth restrictions.

7. Now came my first shock. After exactly 7 minutes and 13 seconds as per my stop-watch (7:06 as per the Safaricom modem) the modem unceremoniously disconnected. My first reaction was that I had been robbed, but checking my balance I noted that I still had KSh. 6.06 in my account. This meant that I had used exactly what was advertised.

Now to analyze this tariff let's look at what I was able to do in those seven minutes:

1. I downloaded 3% of my torrent (appx. 21 MB). This means that to download the entire movie (700MB) I would need KSh. 473.00 worth of airtime - compare that with the 600MB bundle that would cost me KSh. 1,000. Another way to look at this is that the speeds are 50.47 KBps per second on average. (BitTorrent showed a maximum speed of 114 KBps)

2. My Youtube video only played for 33 seconds (11%), the rest of it did not buffer. This represents about 9.4MB since the file is 81MB in size. In speed it translates to 22.78 KBps per second.

3. I was not able to send my attachment.

4. I managed to to update Windows

So what does this all mean:

If you want to use this tariff for downloads (which is what most of us are planning to do), I caution you. The disparity in the torrent and youtube downloads 50Kbps vs. 22Kbps respectively means that you can never really tell how long it would take. Differences in the server you're downloading mean that you will never have the same speeds. I should note here that the torrent file I was downloading was the most seeded file, so you should expect other torrent files might be slower. It also goes without saying that using this tariff with anything other than full 3G connectivity would be expensively unwise.

The 7 minute disconnection was also annoying, since if you are in the middle of an attachment it might mean that you would have to start all over again, losing precious minutes. I am not sure however if this is a problem with the tariff or Safaricom's data service generally.

Later on I connected the modem to see how much I could download with the remaining 3 minutes of data airtime. I managed a miserable 7MB before it expired.

My Verdict

Certainly an interesting way to approach selling data and will most probably bolster revenues affected by voice price wars. I feel however that we moved from per minute data billing (remember Telkom) for a reason and that at the current speeds, this new tariff does not add any value to the data packages currently in the market.

Your tests?

If you'd carried out your own tests with this tariff I'd be happy to know the results.