A partial overview of Ovi Store at 6:30 GMT 30/10/10 browsing it from computer

Hi OviStewey,

I've been in contact with some of Ovi Publish officers since I wrote the last post on the 5th of January. I was informed on the 7th of January that there was a delay in data synchronization since 1st of January. So the list might have seen working well since early December but actually haven't been.

I can give an example of an app developer who feels that is not being rewarded for his efforts. Me. Currently I'm waiting for response to two emails that I wrote to Ovi Publish pointing out that my app is not where the sorting process claimed by Ovi Publish, would locate it.

I've sold more than one item for N86 8MP over the last week and its position is 669 (page 56 on Desktop browser) in the best sellers list. It's fair to say that after some emails to Ovi Publish after the 3rd of January, my app started to have a reasonable position in the list; between page 1 and 10 (even with just one item sold in the previous week ), and sales rose but since about a week ago all dissipated.

Thank you.


[IMG]http://img338.imageshack.us/img338/6193/11012400positionofcount.png[/IMG]

OviStewey suggests, if I'm not mistaken, that Best Sellers calculations are based partly on previous week sales, but balanced with some shorter term sales, which if the case would explain some unexpected results sometimes. But I don't think we will ever know the exact way it is calculated.
It is the second time that a game of mine appears in the top 12 for one day or two, on Nokia N8 , which you think should represent a quite important number of sales, however it was never really reflected in the reports on those days, so either all sales on OVI are very low and a difference of a few extra sales on a day can really have a huge impact in the ranking, or there are other variables that come into play to create a turnover.

I agree it may be unfair if you find your app/game to be always below your supposed ranking. On the other hand if you are in the best sellers but this doesn't reflect any increase in sales (at least not from the estimates), it is more like raising ones hopes in vain.

[QUOTE=quarkos;812985]Hi OviStewey,

I've been in contact with some of Ovi Publish officers since I wrote the last post on the 5th of January. I was informed on the 7th of January that there was a delay in data synchronization since 1st of January. So the list might have seen working well since early December but actually haven't been.

I can give an example of an app developer who feels that is not being rewarded for his efforts. Me. Currently I'm waiting for response to two emails that I wrote to Ovi Publish pointing out that my app is not where the sorting process claimed by Ovi Publish, would locate it.

I've sold more than one item for N86 8MP over the last week and its position is 669 (page 56 on Desktop browser) in the best sellers list. It's fair to say that after some emails to Ovi Publish after the 3rd of January, my app started to have a reasonable position in the list; between page 1 and 10 (even with just one item sold in the previous week ), and sales rose but since about a week ago all dissipated.

Thank you.
[/QUOTE]

You should take into consideration also that the ranking depends on your location. The list (ranking) is looking diferent in London than in Rio for example. (No matter if you browes from phone or desktop.)

Talking about not being rewarded for efforts ... I bet you can find a lot more developers in this situation, myself I have complained a couple of times (or maybe more ;) ) clearly (I hope) expressing the insatisfaction that in our case only about 10% of the development effort is covered by Ovi revenue. But the problem is not in the ranking, or at least not the main problem.

In my opinion the problem is the overall Ovi Store sales compared with the number of items and developers. In the past Nokia and Ovi Store constantly put their efforts in atracting as many developers as possible, but little effort in providing the support for a viable ecosystem. In other words, I do not care if my application is #1 or #99999 as long as my costs are covered and I can grow my business. In turn this will lead to more applications for Ovi Store and will atract more developers resulting in higher copetition and higher quality for applications.

Instead, I guess, Nokia wanted to skip some steps, and they focused on atracting hoards of developers, the result is an amorphous soup of applications where the end user has hard time finding relevant content and where capable developers get drowned. In this situation I tried to survive cuting costs to minimum, obviously this has an impact on applications quality and this creates a vicious circle (less buyes on Ovi).

I think Nokia has understood the problem and they are trying to fix it. What they can do is to promote as much as possible Ovi Store and increase sales. And secondly to provide a fair environment with equall chances for all developers. And here you have a point. If their ranking and promotion efforts are not totaly objective and correct then their efforts my be in vain. But again ranking in best seller list must be corelated with sales in a particular country and this is very difficult since the number of sold items is usually very small, so for the moment I can not tell if it is correct (computer generated :) ), or deceitful (edited by humans).

[QUOTE=intellsys;813246]
Instead, I guess, Nokia wanted to skip some steps, and they focused on atracting hoards of developers, the result is an amorphous soup of applications where the end user has hard time finding relevant content and where capable developers get drowned.
...
I think Nokia has understood the problem and they are trying to fix it. [/QUOTE]


I would like to agree with you, but my opinion is that concern for developers comes last in the list. And this is not specific to OVI, but most app stores (if not all). It is a business, and whether third party publishers are having it good or struggling does not affect their business, or maybe only in the very long term.

[QUOTE=intellsys;813246]You should take into consideration also that the ranking depends on your location. The list (ranking) is looking diferent in London than in Rio for example. (No matter if you browes from phone or desktop.)...[/QUOTE]

I thought on this possibility and went to check it out. Using an Italian proxy server I accessed the Ovi for Italy and took an image of the first 10 pages of the best sellers list and compared it with the first 10 pages of the UK list. They turned out to be identical:

[URL="http://img843.imageshack.us/img843/1928/11012520panorama.jpg"]Comparison of the Italy and UK best sellers list.[/URL]

[QUOTE=quarkos;813324]I thought on this possibility and went to check it out. Using an Italian proxy server I accessed the Ovi for Italy and took an image of the first 10 pages of the best sellers list and compared it with the first 10 pages of the UK list. They turned out to be identical:

[/QUOTE]

[URL="http://img254.imageshack.us/img254/5717/11012600panoramaovibest.jpg"]Comparison of the Best sellers list of Germany and UK for the E72 using proxy server. [/URL]

It may differ between groups of countries, I browes from Romania and it looks different starting from the first page:
[IMG]http://www.insytv.com/extras/best seller.PNG[/IMG]

[QUOTE=intellsys;813396]It may differ between groups of countries, I browes from Romania and it looks different starting from the first page:
[/QUOTE]

Now UK best sellers is the same. I reckon it was updated.

[URL="http://img64.imageshack.us/img64/308/11012601firstpagebestse.jpg"]First Page Uk best sellers.[/URL]

[QUOTE=quarkos;812985]...
I've sold more than one item for N86 8MP over the last week and its position is 669 (page 56 on Desktop browser) in the best sellers list. It's fair to say that after some emails to Ovi Publish after the 3rd of January, my app started to have a reasonable position in the list; between page 1 and 10 (even with just one item sold in the previous week ), and sales rose but since about a week ago all dissipated.

Thank you.

[/QUOTE]

It seems to be fixed, I just wonder for how long.

This makes sense now:

[URL="http://img38.imageshack.us/img38/8069/11012604countdasecondpa.png"]Best Sellers list (UK) for E52 , CountDa in page 2.[/URL]

[URL="http://img231.imageshack.us/img231/866/11012603countdafirstpag.png"]Best Sellers list (UK) for 5630 Express Music, CountDa in page 1.[/URL]

[URL="http://img204.imageshack.us/img204/7927/11012602countdasecondpa.png"]Best Sellers list (UK) for C5-00 , CountDa in page 2.[/URL]

[URL="http://img213.imageshack.us/img213/6303/11012601countdafirstpag.png"]Best Sellers list (UK) for N78 , CountDa in page 1.[/URL]

One item sold for each model in different countries from uk.

[QUOTE=RomanAge;813285]I would like to agree with you, but my opinion is that concern for developers comes last in the list. And this is not specific to OVI, but most app stores (if not all). It is a business, and whether third party publishers are having it good or struggling does not affect their business, or maybe only in the very long term.[/QUOTE]

The "very long term" might be as short as a couple of years.

[I]"It took 38 years for radio to get a market of at least 50 million users; it took television 13 years to achieve 50 million users; and once it was open to the general public, it is estimated that it took just four years for the internet to achieve 50 million users."[/I] ([URL="http://www.elon.edu/e-web/predictions/early90s/internethistory.xhtml"]article here[/URL])

In my opinion in no more than a couple of years we shall have only one major player in the mobile operating systems game. This happens when the majority of the consumers will start using smartphones. The dominant OS will not be Apple's. iPhone probably is a great device but can not cover the entire market, is one device only, same for RIM, not everybody is a business person. So it remains Nokia's OS and Android.

I said it before, Stephen Elop confirmed the imporntance of developers in his short speech at Nokia World in September, so they are aware of it. It will be plain stupidity not to act accordingly.

[QUOTE=quarkos;813417]
...
One item sold for each model in different countries from uk.[/QUOTE]

Correction 2 sales for E52.

[QUOTE=quarkos;813417]It seems to be fixed, I just wonder for how long.

This makes sense now:

[URL="http://img38.imageshack.us/img38/8069/11012604countdasecondpa.png"]Best Sellers list (UK) for E52 , CountDa in page 2.[/URL]

[URL="http://img231.imageshack.us/img231/866/11012603countdafirstpag.png"]Best Sellers list (UK) for 5630 Express Music, CountDa in page 1.[/URL]

[URL="http://img204.imageshack.us/img204/7927/11012602countdasecondpa.png"]Best Sellers list (UK) for C5-00 , CountDa in page 2.[/URL]

[URL="http://img213.imageshack.us/img213/6303/11012601countdafirstpag.png"]Best Sellers list (UK) for N78 , CountDa in page 1.[/URL]

One item sold for each model in different countries from uk.[/QUOTE]

The best sellers list seems to be failing again.

This is an image of my item ("CountDa !") in page 61 (position 723) in the best sellers list. This is from Spain, for the model 6710 Navigator. I've sold for this mobile in Spain in the last five days. It seems that the listing stopped working after the weekend.

[URL="http://img156.imageshack.us/img156/8651/11020100positionofcount.png"]Image[/URL].

I do not like the idea of discussing who's application is the best and who's application deserves to be on the first place a.s.o. I still believe that Nokia and Ovi must focus on increasing overall demand for applications and give all applications equal share of visibility (or at least visibility based on application classes).

But I must admit I have fully understood your point of view today when I noticed one of the applications dropping further in best selling lists while its sales increased significantly due to a review. Seems like there is no correlation between sales and ranks. But then, and here my frustration came in, why do we see same applications over and over again on the recommended page and on the first places in the best selling lists.

Is Nokia not playing fair for all developers? Why doesn't anyone comment on this? How comes that nobody is able to tell us what algorithm is behind best selling lists? If that is the truth than they are surely diving into the wall.

Although under these circumstances the above are resonable questions to ask, I do not believe Nokia mystifies ranking, instead I suspect they have a rather bad algorithm and a strong bias due to sales location and low number of sales. So what I suggest is to make public that algorithm, we all are developers, maybe we can help.

To start with I propose the following simple algorithm:

1) Based on existing ranking divide all N (the total number of apps) applications in M classes of K applications (e.g. K=100)
2) Rotate evenly each of the K apps in one class (each class gets its "visibility window" based on class rating )
3) At the end of a Period (e.g. Period = 1 week) every first L applications (L < K) from class [B]m-1[/B] are promoted to class [B]m[/B] ( m = 1 to M) and last L applications from class [B]m[/B] are downgraded to class [B]m-1[/B]

For transparency and to help developers construct their strategy, publish class ratings outside Ovi Store (not to influence buying decisions too much). You can even restrict republication of these ratings or use NDA agreements. That is just an idea, if you can explain us why such an idea does not work, maybe we can understand better Ovi Store problems and leave aside unnecessary questions.

[QUOTE=intellsys;815597]I do not like the idea of discussing who's application is the best and who's application deserves to be on the first place a.s.o. I still believe that Nokia and Ovi must focus on increasing overall demand for applications and give all applications equal share of visibility (or at least visibility based on application classes).

But I must admit I have fully understood your point of view today when I noticed one of the applications dropping further in best selling lists while its sales increased significantly due to a review. Seems like there is no correlation between sales and ranks. But then, and here my frustration came in, why do we see same applications over and over again on the recommended page and on the first places in the best selling lists.[/QUOTE]

Hi, I do not like either to discuss who has the best application. I would be glad if Ovi Store was a fair environment for competition. As we have seen, it is not. If they are offering this best sellers list, based on the sales of last seven days, then it should be working correctly and not only for one week or two. What is doing currently is to mislead customers (as you will see on many reviews of disappointment of items in the first places) , a real disaster for Nokia and for us developers who are devoting our time for this platform.

[QUOTE=intellsys;815597]Is Nokia not playing fair for all developers? Why doesn't anyone comment on this? How comes that nobody is able to tell us what algorithm is behind best selling lists? If that is the truth than they are surely diving into the wall.

Although under these circumstances the above are resonable questions to ask, I do not believe Nokia mystifies ranking, instead I suspect they have a rather bad algorithm and a strong bias due to sales location and low number of sales. So what I suggest is to make public that algorithm, we all are developers, maybe we can help.[/QUOTE]

The algorithm seems to be explained. The problem in my view is that it is not being used and apps are being located on default positions that change as they add more content to the list.

[QUOTE=intellsys;815597]
To start with I propose the following simple algorithm:

1) Based on existing ranking divide all N (the total number of apps) applications in M classes of K applications (e.g. K=100)
2) Rotate evenly each of the K apps in one class (each class gets its "visibility window" based on class rating )
3) At the end of a Period (e.g. Period = 1 week) every first L applications (L < K) from class [B]m-1[/B] are promoted to class [B]m[/B] ( m = 1 to M) and last L applications from class [B]m[/B] are downgraded to class [B]m-1[/B]

For transparency and to help developers construct their strategy, publish class ratings outside Ovi Store (not to influence buying decisions too much). You can even restrict republication of these ratings or use NDA agreements. That is just an idea, if you can explain us why such an idea does not work, maybe we can understand better Ovi Store problems and leave aside unnecessary questions.[/QUOTE]

I think that the idea of the best sellers list is to promote items that had proven to be popular so to increase the sales of the store. But it is clear that there is nobody in charge worried of what being offered and how satisfied customers are.

I would like to quote a reply from OviAndrew regarding the sorting process in Ovi store written the 17th of February. Since then, I couldn't see any appreciable change in the sorting process and the items when sold don't stay near the top of the best sellers list for a week but for about half a week and then returns to the default position.

[COLOR="navy"]On Fri, 18 Feb 2011 03:26:48 +0100, Ovi Publisher Support
<PublishToOvi.Support@nokia.com> wrote:

Dear Carlos,

I have spoken at lengths with the R&D team of this erratic behavior and
unfortunately this is as designed. This is a simple case of the system
being
unable to differentiate the ranking for applications that have low
download
numbers for a certain device/country combination. Please allow me to
explain
the reason behind this behavior.

The ranking system of the Top free/Best sellers is determined by
aggregating
the device/country combination performance(e.g. N97 mini /UK). Since
there are
hundreds of device/country combinations, these calculations result to
being
approximations of the actual performance which determines the
applications rank
in the store.

For higher download numbers (for example, 1000 downloads versus 800
downloads)
this formula can distinguish the differences between the popularity and
rank
them correctly. Unfortunately for smaller download numbers (say ~1 ?
100), the
system is unable to differentiate the performance comparison for an
application
that is doing marginally better than another. As a result, these
applications
fall within the same group rank and for each refreshing of the
calculations,
they will be displayed in the store with no particular order.

There is a larger amount applications at the low tail end of this
distribution
and all these applications fall within the same rank group. If for
example,
there are 1000 applications belonging to this device/country combination
that
fall within this rank bracket, then each day, these applications will be
shown
in random order. As a result, an application can be shown on page 5 on
one day
and may show up on page 80 the next.

We are aware that this would cause behaviours in the store that appear
seemingly erratic and does not accurately display the popularity of the
applications; in the following weeks, we will be putting in place more
rigorous
data monitoring mechanics to ensure better accuracy in the ranking
system and
eliminating such discrepancies between the performance numbers.

I apologize for the confusion and inconvenience caused by these issues.
Rest
assured we are working diligently to enhance the current system so as to
provide more accurate results to our end users. In the mean time, I ask
for
your further patience and kind understanding while we migrate to this new
ranking method in the following weeks.

Best regards,
OviAndrew
Ovi Store Publisher Support
[url]http://publish.ovi.com/[/url] [^]
Promote your Ovi Store App on the web now!
[url]http://promotions.ovi.com/omt[/url] [^]
Ovi Publish Checklist
[url]http://info.publish.ovi.com/?cat=34[/url][/COLOR]