First of all I'd like to commend the developers of this site for undertaking such a massive task (i.e. programming custom flash modules for thousands of individual Chinese characters) and creating an awesome study tool for those of us toiling through the process of learning one of the hardest languages in the world. In just a few short days of using the site I've definitely noticed an impact on my writing abilities, i.e. stroke order mistakes that I've never been aware of in the past are getting slowly corrected. I think it's fair to say I'm addicted!

Now then, I do have a question and some suggestions for the developers:

Questions:

1. I have a single list of vocab that I'm drawing from at the moment for practice, and I've set the Add Word Frequency to "Early and Often" (after looking for a loooong time and finally finding this setting in "account settings" of all places). However when I go to practice, I only get a set of about 8 words which constantly repeat themselves, even if I get them all right I still get the same set of 8 words unless I manually refresh the page, in which case I just get another 8 words. I assumed that the software would look at which words I get right and naturally swap those out for new words that haven't yet been tested or haven't been tested in X amount of time? How exactly does this whole process work? How many times do I have to write a word correctly before it's "learned"? To me, if I write a word correctly the first time it should be automatically counted as "learned", but that doesn't seem to be the case. I read through the entire site for info on this and found no basic description of exactly what is supposed to be going on during the practice sessions. In the end I'm just getting frustrated without any kind of real feeling of progress.

I tried changing to "manually add words to practice", assuming that the little green plus sign would let me add new words to my practice queue, but it simply goes gray after starting a session. The "X to review" number seems to jump around randomly, and I see no way to find out exactly where this number is coming from. Additionally the page is showing "added 4" constantly, but which 4 words were added? Every time I refresh the page I get 8 or so different words. And again, no matter how long I write the same 8 words, I never see any new ones without reloading.

For what it's worth, here's how I envision the learning process would work best, at least for me:

The software looks at my entire vocabulary pool and generates a score for each word based on percentage of times the word was written incorrectly (no impact if it has never been written), and the amount of time elapsed since the word was last tested. As I continue practicing, the software automatically draws from my vocabulary pool in order of the aforementioned score. In this way I can write as few or as many characters as I want in a given session and always be guaranteed to be testing the characters I need to learn the most at any given time. It would be helpful and easy at this point to offer some kind of "meter" that shows where the word I am currently writing relatively fits in this score range. If I'm on a "high-score" character that I've written incorrectly 60% of the time, and hasn't been written in 2 weeks, it's safe to say there would probably be many other "high score" words in my queue, and the meter could display as "full", meaning I have a lot of work to do to lower the scores of these relatively difficult words. If I keep up with my reviews, and the current word I'm working on is one that I've written a week ago and written correctly 95% of the time, the meter would read closer to "empty", as there are fewer words in the pool that badly need to be reviewed.

The next point I want to make has to do with the price setup for the site. In my strong opinion, $9.99 a month is far, far too expensive and I truly believe this pricing scheme will turn away a great number of customers - in the end I think you will lose money by charging this much. Your learning system is intended to work over long periods of time - months, and for those of us who stick with learning the language, years.

By setting a montly fee instead of a one time flat rate fee, as a potential customer I immediately start calculating in my head - $120 a year.. $600 for 5 years (the average time it takes for someone to get decently fluent in Chinese). Ouch.

Undoubtedly your software took immense time and resources to develop and you should be rewarded for that - but $600 for 5 years of usage is downright scary. Personally I would feel a lot more comfortable paying a one time fee of $50-$100 for a lifetime subscription to the service. But if you insist on a montly price scheme, $3-5 a month would be a much easier pill to swallow.

Look at the Iphone games market as an example - the top developers are making millions by selling games for $1 to very high percentages of Iphone owners. You could do the same - become the indispensible tool for Chinese learners by making the price so low you'd be stupid not to use the service. The question of course is, at a price point of $5 a month does your market increase by at least 2 times? Yes I think it absolutely would. Where the optimum profit line lies if you graphed this out is impossible to know but I can tell you it is not at $9.99 a month.

Back to the one-time payment pricing model, one benefit to you would be not having to worry about competetitors taking your current customers - if I'm only on a monthly pricing plan, and a competitor comes along with a better deal, I'll jump ship right away. On the other hand, if you've already gotten my $50 up front, whether or not any competitors come into the picture later is a moot point, at least as far as your established customer base is concerned. Your first in the race to create a learning tool like this and if I were you I'd be interested in getting as much up front as possible.

The other benefit is exploiting the vast majority of learners who, at the beginning of the learning process, have great expectations for themselves and "know" that they'll stick with the program for years, making the up front payment of $X easily worth it, and 4 weeks down the road realize they aren't as interested as they thought they would be. This is why most gym memberships tend to be based in years rather than months.

Obviously it's your business to run but I think you have incredible potential with this site so I wanted to make clear my feelings on this in detail. So enough with that. :)

One other point - you definitely need to make the "help" section much more prominent on the site. Right now the way everything is set up, it is very difficult to figure out exactly how to use the site, and the only way to learn about how to use it is to click a tiny help link on the bottom of the page. Even then most of the information is outdated and seemingly pertains to an older version of the site. At the very least I would suggest having a basic getting started page that goes through the basics of adding vocab and various settings for practicing, explaining how everything words as opposed to simply detailing what various options mean.

Whew - that's all I've got for now - thanks for reading/helping!

Hi

Skritter should work more or less how you describe you'd like it to work. You shouldn't just see the same 8 characters (or a different 8 characters all over again).

As long as you're in normal practice mode (and not cram mode), once you've learnt the characters you should have new ones being added. That is assuming, you've selected a list - which you have, are automatically adding from it (check the settings tab on the practice page), and assuming you haven't come to the end of your 2 week trial or somesuch (if no subscription, new words can't be added).

Do double check you're in normal practice mode, not cram mode though...

As for pricing, the system used to be like that, but they changed it to a monthly one a few days ago. Jury's out...

(BTW the list selected should be 'added to normal practice' or somesuch, rather than 'practice as a cram list').

Thanks Byzanti for the ultra fast response! I am definitely in normal practice mode - as for having a list selected, this is a confusing part - I initially add a list, and it shows up in the active lists section. If I refresh the page, however, and go back to the "active lists" settings area, that list no longer shows up in the active lists section, which is now empty. The vocab words do still show up in my practice queue (as I only have one list at this point).

So I assume once all of the words are added from an active list, it gets removed from the "active list" page?

The weird thing is.. the list has 79 words. Right now in practice mode it shows "65 words to review" and "added 4". That first number seems to keep going up periodically. Shouldn't it start at 79, and slowly decrease as I have fewer words left to review?

I just spent half an hour practicing, it sticks with the same 8 or so words indefinitely. I checked my vocab queue with another browser window as I went along and it doesn't seem like any of the scores are being impacted, a couple of the words are fully learned as well but continue to reappear in the small bunch of words. Any idea?

For a more concrete example, I just reloaded the practice page, and the first item that comes up is "迫不及待", a word that I've already learned according to the vocab detail page, and this word is supposed to be 9 years away from coming up again. On the practice page, it lists this word as being studied -257956321 seconds ago, not due (70%). This appears to be a bug?

Ahh.. I think I figured out part of my problem - I turned on the grading buttons, but I wasn't grading the ones I answered correctly - I only marked the ones I didn't know as incorrect. This must be my problem!

(Although I still don't know what's going on with the aforementioned bug)

Yeah, the not due (x%) is a bug I've been getting too a bit recently. I usually refresh the page if that happens.

As for how reviewing works, lists you've selected add new words to a master list of your own. So "64 to review/4 added" means that Skritter thinks there are 64 things already in your master list that you should be reminded of (words, or tones, or definitions etc), and that today you've added 4 new things.

So, once you whittle the 64 down a bit you should start seeing the "added 4" bit increasing as new words/definition prompts etc are added to your master list as you have less to review. So it wont add the 79 all at once (and if you're studying 4 parts - tones, writing, pinyin and definition, that's in total over 300 new bits (at 4*79).

If you want to see new words earlier, you can change how often they're added in settings, or press the + button...

Okay, apologies for talking to myself, but there's definitely something fishy going on. I am having to write a set of 8 words about 20-30 times before I get one of them switched out for another word on my list, despite the fact that in my vocab queue these words constantly showing up have a "Next" value of at least 60 minutes. All of the other 70 words in my list show as "Ready" in the Vocab queue, but take forever to actually show up.

I tried deleting all of the words in this list and reimporting them, but the server seems to keeps my data for these words anyway. I can't get rid of two words from showing up every single time, both show up as having been studied negative seconds ago, such as "-257956321" seconds ago.

In that case I have no idea!

Yeah, unfortunately every time I refresh the page I'm still getting those two same phrases in the "rotation of 8".

If I'm understanding you correctly, I'm not sure why it says 4 added when I just added the entire list of 79 words in one shot today? Is there some kind of middle layer going on I'm not aware of?

If it currently says "49 to review", I should be seeing all 49 characters being randomly chosen throughout my practice session without having to refresh the page, correct?

Thanks for your help Byzanti, VERY much appreciated. Scribble is proving to be as difficult to use as it is powerful. :(

Sort of. Those 79 new words will need added to your general practice - the pool of characters you practice from - before you'll see them. They wont add if you've still got a load to review.

The '49 bits to review' means that there are 49 prompts remaining till Skritter thinks you don't have any more due. But those are from characters already added. Not new ones from your list. On your current setting, those will get added when you have maybe 5 left to review.

Of course, if you've been getting the old characters wrong over and over again, they'll keep on appearing and you wont get many (if any) new ones. Otherwise it's a bug, and someone from Skritter will need help you there...

Hmm, sounds like a Flash problem, and our Flash developer has just gone to France for a week! It's not good though, so I'll try and debug it as best I can.

I've checked the backend, and my best guess is it's timezone related. What timezone are you in? Have you set your timezone on the site correctly? And is the timezone on your computer set correctly? There was a bug Nick fixed a while ago, trying to make sure the Flash did not rely on the computer timezone, but maybe he missed one. But that negative value for how long ago something was studied is good evidence that the timezone is the problem...

I'll just post this so you can see it now and write another post about pricing and help documents.

Thanks Scott - yes that could be it, I'm in Taiwan, and I never set the timezone on the site - I'll do that now.

Okay - I've got the timezone set now, from what I can see I'm still getting the same bug. In my case it's "迫不及待“ and "強悍“ that are the culprits.

Thanks for your help Scott.

In terms of pricing, it's very hard to say what is optimal. It's a complex problem, and we only get so many shots to try out different methods, because we don't want to be constantly switching around our price just to see if any one change will be for the better. We can hypothesize all we want what people are willing to pay, and come up with different systems, but in the end testing those hunches is really the best way to see and then acting on the results. And we can only test so much.

As Byzanti mentioned, we only just recently started this month-by-month system; before we had it the way you describe, with users purchasing lengths of time whenever they choose, and the longer the time length, the lower the price (as low as $6 per month if you buy two years worth, the highest being $11 for one month). We used that system for 8 months. We know how well that system works. It works okay. But it does have its own downsides.

So now we're gathering data on this new system. It's too early to say with certainty if it will win the customer wallet vote, but so far the sky hasn't fallen. We'll give it time, and then we can determine what's better, the prepay system or the month-by-month subscription system. And then we'll go from there.

We don't want to talk too much about our pricing plans and thoughts, but I can say that we're thinking about adding an educational discount, so that anybody with a .edu email address can get a discounted price. We've found that students often don't have a lot of extra cash for things like skritter, so we want to see if we can find a price that makes sense for both them and our business. We'll see!

Also, bear in mind that you can continue to use Skritter even if you're not paying. So if you study for a year or two, you can still use Skritter to keep fresh those characters and words in your mind without having to pay.

Thanks for your thoughts on the pricing though, it's really good to get some frank thoughts from our users on it. Pricing is something we need to get right, so we do take everyone's opinions into account and spend a lot of time thinking about it!

Ack - indeed the time on my computer was the problem - my little tablet PC I am using was set to 2002! After correcting that everything is working perfectly. Doh :(

Aha! Well then Nick has a bug to fix when he gets back... that would explain why many of your items are marked as having been studied in 2002. Ideally Flash shouldn't look at your computer time to determine anything, but it must be getting it from somewhere...

Glad that's figured out! I'll see if I can hack something together so this doesn't happen until Nick returns and properly fixes it.

Ah yes, and your third and final point: the help system. Writing a bunch of comprehensive articles on how to use the site has been on George's to do list for some time, and I just prodded him to step it up (in all fairness, he has about a couple dozen super high priority things to do). He says he'll work on them this afternoon. That's something that really should get done soon, we agree!

My 2 cents...$10/month is a fantastic deal and I'm happy to pay it. There aren't many good alternatives out there, certainly not many cheaper. If the price is driving potential customers away, I have to think they're not really too serious about learning the language. I wouldn't be too worried about losing customers who aren't "locked-in" by a one-time, lifetime membership to the competition. If they think there is something better out there, having paid a relatively low one-time fee at some point in the past is not likely to keep people from jumping ship. The very nature of the tool means there is an incredibly high switching cost, though - you essentially have to start over again at zero. You lose all the history of what you've studied, what you know/don't know, you lose all the time you spent building custom word lists, etc. A new tool would have to be incredibly spectacular to make me want to do that. Not too likely.

As soon as a truly competitive market develops, the price should approach marginal cost, i.e. a user pays as much as his usage adds to the amount of total costs.

The most expensive part of a service like Skritter is fixed costs (development of the technology), marginal costs should be rather low (although I assume that costs for the usage of the Google App Engine aren't insignificant).

The key question to ask is whether you have enough of a head start to keep your monopoly position. Given that you have to develop a user base from scratch and initial investment is quite high, the ROI for potential competitors is uncertain. Thus, I assume we won't see true competitors any time soon. This could change in a more mature market.

Although I am a "heavy user", $10 are above my limit. I'll either purchase some 'time' at the old prices or pay every X months. That's simply because I'm a student, so I'm waiting for you to introduce the suggested 'price discrimination', i.e. lower the price for students and thereby expand the customer base.
Price discrimination isn't necessarily bad, it's completely rational behaviour of a price-setting supplier.

I am a rising economist (a junior in college) here, and I agree with the post above. Price discrimination can help you guys get more of the rents out of the dead-weight loss (from people who can't afford the service at this price, and from YOU SKRITTER, who can't collect rents on the the people at the margin)

You guys should significantly discount the price for students. Please set it like less than 25 dollars if you subscribe the whole year or $5 a month. I am sure you will get a lot more subscribers than you do now.

Many of my friends are interested in using Skritter, but are thrown off by the high price.

Not to mention, and I know you know it too, Skritter is buggy.

I am not an economist, but as much as I'd like to pay less for Skritter, I also like to see them have the money to improve things :). I think one more factor to consider is that the Chinese studies market is much smaller than, say, the iPhone games market. Therefore, a $1 strategy is much more dangerous to Skritter than to an iPhone game. $10 is not little money, but with the kind of response Skritter is getting, it might help to build up some solid funding. It sounds like a pricey strategy 37signals would use.

I'm not convinced lowering the price as significantly as many people suggest would bring in significantly more customers... bare in mind if you halved the price per person you would have to double the amount of users to continue to generate the money.

For services like this, I totally believe the phrase "you get what you pay for", and even though, as a part-time student/volunteer worker, i can't really afford it, I'm more than happy to pay $10 a month for the service (heck, using it has got me to the top of me Chinese class). The user base isn't huge, and i would much rather pay that much and see Skritter constantly updated and improved than see the programmers become unable to devote their time to Skritter because they had to go do other things to survive...

That said i would like to see the discount for bulk purchases return...preferably with paypal payment as an option :D

From my understanding, the bulk-at-discount payment method is being taken away because it leads to confusion...?

Why not reorganize the payment page instead of taking away the reward for long term commitment?

We all know motivation helps, a lot. As a teacher in China making RMB, it's a nice chunk of money to pay for a year of Skritter, even at the discounted price. Buying a year subscription means I'll be coming back again and again to use it, not only because I like it, but because I want to get wait I paid so much for.

For many reasons: Reward those who make the commitment. Any long term commitment discounts you give will pay for themselves many times over through word of mouth and return customers.

It's not for me, I'm buying a year or two in a few days, it's so Skritter is around in 3, 5, and 10 years down the road.

I also liked the pay X time in advance options and I wish it had not changed.

I definitely enjoy not having to go through Google to pay though. Google is Evil.

To me $9.99 is spot on. However, it seems I am one of the few people using this site that actually has a paying job :)

Some random thoughts:

I wonder what if Google the company goes bankrupt will Skritter still work? Can it be ported to something else? Will the cost of doing so be prohibitive? What if Nick leaves will Skritter still continue?

By the way when I started out with Skritter I found it addictive and easy to use, although it took me some time to get things setup perfectly.

yeah the so-called "buginess" of Skritter is not something I have experienced with any regularity over about a year of use...

@skritterjohan if Google goes bankrupt, I think there will be some bigger problems than Skritter becoming inoperable... also I feel that Google would allow the Skritter team to access their data in a meaningful way, but I have no real experience with the App engine.

I also have a (low) paying job in China but of course find Skritter to be quite worth it. Sure students might not have as much cash, but if they have the time to study Chinese and want to make it a priority, $10 is a fairly small further investment (and one that I find parents are sometimes happy to help with). I mean we are talking like 2 cups of coffee, 1/5th of a new video game price, 1 movie ticket, etc.

I'm not saying there aren't people who can't afford 10 dollars a month, of course there are plenty around the world.... but out of those, the people who have the interest and time (and internet access)needed to study chinese and need this software seem quite few.

Google going bankrupt would take a while, so we'd migrate off before App Engine turned off. They've guaranteed App Engine at least three years' notice and I doubt it will ever shut down.

I don't think I could ever fully leave--Skritter is my baby. Even if it grows to the point where it doesn't need me any more, I will still probably try to meddle with it, like an outgrown parent trying to come along on trips with his kid's friends.