Sprint Customers: normal service will resume soon

WAP Review first brought the news that Sprint's transcoding technology from Openwave was unable to discern in a large number of cases whether a destination site was already mobile or not. The result being that a number of sites would automatically be garbled when in fact they had already been configured for mobile. Sadly, this includes Mippin.

There is a lot of debate about how to improve the intelligence of the transcoders (or more particularly the intelligence of those implementing them as Sprint's issue is a carbon copy of the much discussed Vodafone implementation with Navarra last year) but in the meantime to resolve the issue you need to request to be put on a white list that the transcoder will not touch.

To this end, Sprint set up a page in its developer site to allow people to register their requests for whitelisting. Thanks to WAP Review we are in the first cluster but there are now pages and pages of requests from a host of developers. The current promise is to have the issue resolved within 30 days - I suspect there may now be some slippage.

If you are a Sprint customer using Mippin, please hold tight. We do have a fix in place should Sprint not resolve it quickly but we really do not want to use it. We do share your frustration but it will be back to normal soon.

A cautionary SMS tale: Part II

Note to self:

It is not enough to remove the link for the ability to share via SMS; we must also remove the actual page from the production servers. Some people apparently bookmarked this page and have continued to make hay.

For those new to the topic and for additional context: click here.

Keep your eyes peeled for the next installment. We think that it is finished now. But we have said that (many times) before.

Oh... and the reason for this picture? The Mermaid Poker Challenge (which you can actually download for free if you wish) is now one of the most popular stories within the Mippin service having been "shared" many many times.

The long tail in action

Hats off to the staff at this junior school - unknown to us previously - who have seen Mippin and used the Mippin Maker to ensure that parents and children can keep up to date on the school's events even when away from the PC.

A true example of the viral spread of Mippin and the ease with which any publisher can get up and running. We hope that many others, from many different walks of life, follow the same example.

Note the Make It Mobile button also on the site. An easy addition accessible from the Mippin Maker and one which really makes a difference in increasing take-up of the service.

More metrics

AdMob is certainly one of the movers and shakers in the mobile space and, in our experience, staffed by some of the nicest and most knowledgeable people in mobile.

The latest monthly metrics are out and can be found at www.admob.com/metrics. Always worth a scan.

More deep diving on devices.... and the winner is?

We have taken a deeper look at the devices we are seeing come into the service and how effective they are at generating repeating active users. We used this data also in exploring our cookie conundrum and as a result we might see some changes in the relative effectiveness of the different manufacturers now we have put a fix in - however there are still some nice observations to be made.

The tables below highlight the market share that we observe in each market for each major manufacturer. But also the relative effectiveness of each device in each market in generating a repeating active user. Thus for the latter the top rated device in each market is rebased to 1 and each manufacturer is rated according to that device class. Thus with a rating of 0.50 we are saying that the device is half as likely to generate a repeating active user than the market leader.

All good? OK.

First the UK:
The headlines here are that Sony Ericsson leads the market amongst mobile data users followed closely by Nokia with the others far behind.

However, when it comes to creating active users LG and Motorola lead the way. This appears somewhat counter-intuitive but for additional context one needs to bear in mind that a significant proportion of users are on the Three network which has the most active user base in the UK. Samsung is clearly a concern in this summary and may be related to our recent cookie investigations into particular gateways and device browsers. If so, we should see a marked improvement in this stat in the next weeks.

Now, South Africa:
In South Africa Samsung lead the way followed by Nokia and then a drop down to Motorola and Sony Ericsson. The most effective devices are the LGs and Sony Ericsson's with Nokia actually being a bit of a laggard and as we progress you will actually see this as a persistent theme. We believe this is due to the Openwave browser on MIDP 1.0 and later Series 40 devices and again we shall see if this is the case with the recent fix. Certainly, the data is surprising in an industry where Nokia appears to lead in all categories.

India

This exact same trend is also seen in India where Nokia has a significant market lead.

US
In the US, the clear lead of Motorola in volume terms is obvious but the iPhone share of mobile data activity is impressive. More so when you consider that it is also the most effective device in the market. By contrast Motorola performs poorly in these terms. We expect all of the laggards to improve but not to reach iPhone levels. That Sony Ericsson's get close is impressive - just a pity here as elsewhere that its market share is not higher.

As stated in the cookie analysis, some of this data regarding manufacturer performance may create some distortion peculiar to us, but the market shares should be a solid reflection and come from a sizeable database these days. Would be happy to hear of, or compare and contrast with, others' experiences.

Bookmarking Mippin Today

Thanks very much to many of you who have sent us feedback on Mippin Today.

The most consistent piece of feedback is how do you bookmark Mippin Today or what is the URL to set it as your home page. Shame on us: we should have mentioned it in our first post.

Please point your phone browser to mippin.com/today

We are really glad you like it. Keep the feedback coming.

Mippin Today

If you haven't tried Mippin Today so far then please give it a spin.

It is the closest thing to a newspaper written especially for you that you will find - in fact better than that - as it is a newspaper written for you when you want it, whenever and wherever you are.

It brings together the most recent content from your favourites into one place.

We have also added the ability to receive an email once a day with your latest edition.

Please give it a try and let us know your feedback - in the comments to this blog or from the bottom of every page within the service. We read them all.

Cookie Conundrum

This is a post for the techies but one which other developers might appreciate.

Cookies are key to Mippin - we use the cookies to help shape the service around your usage so you do not have to register to get the majority of the features. When those cookies do not appear, the service loses a lot of its personality. It still works nicely but it does not shape itself around you - always leading you back to a generic home page.

Our analysis of three of the most popular markets showed the following ratio of devices not supporting cookies:

South Africa - 4% unsupported
UK - 6%
and
US - 43%

The last number really took us aback and we have spent quite some time querying the data.

It raised more surprises. Usually in these types of analysis Nokia scores the most highly and Motorola much less so. But in this data there were a very high number of early (MIDP1) and Series 40 Nokia that have no cookie support, indeed far more as a percentage of those total devices from Nokia than from any other manufacturer (60% vs 45% for Motorola for example - though the total number from Motorola was around the same given its superior market share in the US - over 2 to 1).

We were able to isolate the instances predominantly to the networks of Verizon, Alltel and MetroPCS and began to wonder if the policies or infrastructure of those operators meant that cookies were purposely withheld. (For example, use of an older version of an Openwave gateway by O2 in the UK withholds cookies).

The apparent answer we found through a simpler means.

Each of the devices were returning detail of the browser in the user agent and for each of those devices using the Openwave browser UP.Browser versions 6.2.3.*.

We found that the cookies were withheld for this particular browser. We looked a bit closer and realised that this only occured when the URL has no sub-domain: a problem when we have been promoting http://mippin.com to cut out on the number of key taps for the end user.

The problem handsets successfully store the cookie with a domain name they understand i.e. .(dot)mippin.com, but when they return to mippin they cannot match the domain .mippin.com to our site mippin.com and don't send us the cookie.

Other browsers deal with this without issue and now we know what the issue is, we can deal with it too. So: for users on a number of Series 40, or the high volume Moto KRZR or RAZR from the US operators above, a remarkably improvement in Mippin will occur.

Phew!

An excellent blogging aid

We wrote an entry at the end of last year encouraging people to write their own blog - there are an estimated 120m blogs worldwide - the long tail in action without a shadow of a doubt.

This week were re-acquainted with Zemanta and it is certainly worth a belated mention.

Zemanta helps you get the most out of your blog, suggesting images, links, tags for your posts: helping bring it to life and extend its reach.

It's a very very clever tool. We tip our hats to them. Give the demo a try.

SMS again: genuinely amusing this time

Check out the video within this post over on Mobhappy. A classic.

An amusing but cautionary SMS tale

Regular users of the service in the UK will have noticed that they had the option to share stories with others through SMS. We trialled this feature in the UK first intending to roll it out globally if it proved successful.

Very vigilant users will have noticed that this SMS share option has disappeared.

Why?

Well, we witnessed some very interesting use cases that had nothing to do with Mippin. This is a long story so please make sure you are sitting comfortably and I'll begin.

In our first iteration we provided the ability for someone to add a message to their friend to give the incoming text some context e.g. "take a look at this.....". However, an example of what we actually saw happening was this:

Example 1

11:36:18,828 - 0044777083xxxx BODY=Your friend cal says: call your mum chloe asap

Example 2
12:43:22,008 - 0044771767xxxx BODY=Your friend Steph says: Ive done it alg emailed debbie this mornin and said that im in level 1 but she said im in level 2 in

Example 3
12:45:24,313 - 0044772763xxxx BODY=Your friend donna says: hi babe x just got back xxx

Example 4
13:02:40,805 - 0044794739xxxx BODY=Your friend Christian says: Yh for 2minz 2 get ma cv, im wit ashleigh nw n i got a job day 2mz frm 10-5 lol

Example 5
13:06:16,968 - 0044794739xxxx BODY=Your friend Christian says: In da 2nd hand furniture shop, i dno didnt c her

Example 6
14:02:02,853 - 0044777450xxxx BODY=Your friend Sadat says: Hi cyn why u didn't tx m chat msn me? Im sad now. I stil horn..

Example 7
17:01:59,308 - 0044751625xxxx BODY=Your friend djoakes says: hi bbe am not comin down just mashed up ma leg a car turned da corner n a mashed up da left door uk

Example 8
18:16:47,436 - 0044783331xxxx BODY=Your friend KITTY says: PAUL CAN I SPEAK 2 U ITS URGENT

My favourite is #7 - surely Mippin wasn't the most effective way of communicating this message. And before you check, I can confirm that the article has nothing to do with the content so it did not add anything to the delivery.

In other words, a segment of the users began using Mippin as a free P2P messaging service despite the inconveniences of doing so. This actually generated over 95% of messages through Mippin.

Now at one level I just want to tip my hat for the ingenuity and we might seem spoilsports by looking to discourage this activity. However, each message actually costs the company money and eventually means less spending on making the experience better for everyone else, so we looked to close the loophole.

The change suggested was to take away the free text entry field and only allow the end user to enter their name (up to fairly short maximum number of characters), e.g. "John has sent you this story..."

Bizarrely we saw the amount of messages increase. Significantly.

Scratch head time.


Then we realised something. These users were not changing their behaviour. They were simply dividing their messages into smaller chunks so that they could use the name field to send their messages bit by bit.

Now, this is fiendish in the extreme. Tedious for the sender and incredibly irritating for the receiver. But this loyal base of free messaging addicts did not appear to think so and we saw the number of messages sent escalate considerably.

Back to the drawing board.

Our last option was to exclude the user name completely and remove all mention. But in this we have stopped short. First, to the end user this would look like Mippin was sending an unsolicited SMS. Second, we noted the growth after the second iteration of one or two individuals using the feature to spam others repeatedly - many, many times.

Thus, dear reader - if you are still there (I told you it was a long one) - : there is no longer the ability to share via SMS on the service. And if you provide such a facility then be aware that a large number of text addicts, formerly with Mippin, are heading your way.