PDA

View Full Version : Full metrics on the muse I'm testing - I'd love some feedback


SoftwareGeek
01-29-2010, 06:32 AM
I haven't seen a ton of numbers and detail about muses, so I thought I'd throw mine out there for everyone's benefit. I'd also love to get some feedback. I think I'm at a reasonable go, no-go decision point, and I'd like to get some other opinions. In the interest of everyone (including me) learning as much as possible, I'll try and be as completely open as I can.

First off, the landing page:

http://www.gradeomagic.com

The service is a grading service for teachers. The idea is that they scan or fax their quizzes, tests, and homework in and mechanical turk workers grade it. Once grading is done, they get sent an email to a complete grading report broken down by student and by class, plus a pdf with graded versions of all the work, plus grading slips they can print out and attach to the original hard copies.

There are existing web-based gradebooks which seem to have reasonable traction(http://www.mygradebook.com is an example), but no service that will actually do the grading for you. After talking with teacher friends of mine, they said they'd be interested in using such a service, and would be willing to pay a decent amount for it.

I've been running adwords to drive traffic to the site for about a month now, and my stats are as follows:

Total adwords spend: $232.85
Total adwords clicks: 284
Total adwords impressions: 39,316
Average CTR: 0.72%

Total adsense spend: $14.28
Total adsense clicks: 20
Total adsense impressions: 50,849
Average CTR: 0.34%

Total facebook spend: $64.72
Total facebook clicks: 61
Total facebook impressions: 497,762
Averate CTR: 0.012%

I've gotten 12 conversions, 10 on the old site (http://gradeomagic.weebly.com/), and 2 on the new site. When I switched from the old site to the new site, the old site was getting around 5% conversion rate from adwords traffic. However, the old site only asks for the user's email address to be added to a beta list (though it's below a headline stating the service costs $50/month). The new site leads the user to what looks like a full signup form with a price clearly at the top of the page, and a full name, address, email, etc form. As far as I can tell, adwords traffic is the only traffic that's converted. Adsense and facebook were total wastes of time and money.

I was rather excited about a 5% conversion rate. However, on the new site the conversion rate is more like 1%. It looks like only 16% of the traffic that goes to the homepage makes it to the signup form, and then only 6% of those people convert.

So, the first big question I have is whether I'm going to be able to drive enough conversions for little enough money to make this fly. 1% conversion rate seems pretty low.

Now, on to cost.

I figure I need to have 2-3 workers look at each page of the test. I spent $150 to have a Romanian programmer write a quick flash app to do the actual grading, and put some sample jobs up on Mechanical Turk. I started at $0.10 a page for a 2nd grade math quiz, which several workers remarked seemed like a really good wage. I then tried another job at $0.01, and got reasonable quality work, but there was less satisfaction with the wage.

I figure that an average teacher might have as many as 80 students. This means if they give a 2 page test (single sided, or one double sided page), then this results in 480 pages that need to be graded, if every page is graded 3 times. The total cost of this, assuming I pay workers the lowest wage possible of $0.01, is $4.80. If I put a 2.5x markup on this, this becomes $12.00. Based on feedback from the teacher that sent me the test quizzes, something of this magnitude would take her 2 hours to grade, which translates to an effective $6/hour wage/cost for her.

I can't imagine marking these prices up that much higher and still getting teachers to go for it. However, I'm really concerned that the markup is too low (Tim recommends 8-10x), and having a rock-bottom worker wage doesn't give me much room to play (doubling it to $0.02 throws the numbers totally out of whack).

What do you think? Is this a viable muse? Is it worth continuing to test, or is it best to throw in the towel? What changes would you test before giving up? Or are the numbers just not good enough to continue?

I'd love to hear any and all feedback you have, and I hope the above numbers are useful to others.

jon123
03-08-2010, 06:02 AM
Right now I believe your conversion stats are statistically insignificant (try http://www.prconline.com/education/tools/statsignificance/index.asp ), but it's tough to say since you didn't specify how much you spent/how many visitors caused the 10 signups vs the 2 signups. You have these numbers though, so give them a try in the calculator.

Regarding your pricing ($50/month), have you considered a per use fee? For example, MailChimp.com does email marketing and they offer a variety of plans - one of them is pay as you go, and you buy "credits", each email costs 1 credit. You could charge 1 credit per exam (or even per exam, per page or some other metric), so if they have 80 students, and it's a 2 page exam, it costs them 160 credits to have your site do the work. You can determine the actual cost per credit this way since you know how much you'll be paying workers to do the work.

There could be other issues with this type of a service as well, though I claim zero knowledge of this, so here is some speculation:
Legal issues: Teacher agreements? Copyright (could students claim their work is copyrighted?)? Privacy (students)?

jon123
03-08-2010, 06:21 AM
I hadn't checked out your site prior to posting - but make a minimal investment to host the site and remove the spammy looking footer: "Designed by Free CSS Templates and Dubai Apartments" and "Make a Free Website with Yola".

Would you give money or sign up with a site that had those on the site? I know I sure wouldn't.

You can get web hosting for $5 or less/month. If you are going to be testing other muse's you could even get a reseller account for probably $20 or less/month that lets you host many many different sites. If at some point a site became popular and required more resources you can easily move it over to a VPS/dedicated hosting.

I personally like to use Wordpress for my sites, and I just modify it to work more like a CMS (remove all links to comment sections, posts, better urls, remove rss feed, etc). There are some themes designed specifically for using Wordpress as a CMS, or you can mod the theme yourself (normally, check licensing restrictions).

jon123
03-08-2010, 06:25 AM
Ok, sorry for another post but I keep thinking of things here, last one.

Maybe you allow them to submit by mailing them to you (someplace that would scan for you)? Or taking pictures with a digital camera? Or provide an "admin area" where they can upload documents and retrieve old docs/results...

I think it's a decent idea, you just need to make it really easy for them to use, and continue to use.

vtamethodman
03-09-2010, 02:14 PM
Put up a free trial button, it's amazing how fast your conversions will go up. Here is my website as an example http://virtualteachingassistant.com

My bounce is approximately 60%
Conversion for free trial is 4% (anybody know how I can get that higher? I'd think more ppl would sign up if it was free)
About half of free trials results in a sale 2%

Mturk is a great service, I actually use them for my free trials!

jon123
03-11-2010, 08:41 AM
How do you use mturk for your free trials?

vtamethodman
03-11-2010, 06:38 PM
If I need a paper corrected I'll export it to mturk and then review it before submitting it to a student. Pretty easy and usually only costs me a buck per paper.

MattInglot
03-11-2010, 09:48 PM
I've gotten 12 conversions, 10 on the old site (http://gradeomagic.weebly.com/), and 2 on the new site. When I switched from the old site to the new site, the old site was getting around 5% conversion rate from adwords traffic. However, the old site only asks for the user's email address to be added to a beta list (though it's below a headline stating the service costs $50/month). The new site leads the user to what looks like a full signup form with a price clearly at the top of the page, and a full name, address, email, etc form. As far as I can tell, adwords traffic is the only traffic that's converted. Adsense and facebook were total wastes of time and money.

I was rather excited about a 5% conversion rate. However, on the new site the conversion rate is more like 1%. It looks like only 16% of the traffic that goes to the homepage makes it to the signup form, and then only 6% of those people convert.

1% is an OK place to start, but now the challenge is to get that number up. Reading your post I think there are several important factors that explain the 5->1 shift:

1) The test site and the actual site are completely different. Your Weebly one may have simply been a higher converting page than the new layout.

2) You were smart to put the price of the service into the test, but recognize that you will still get people interested in learning more about the service signing up with their e-mail, but who would not have necessarily whipped out credit card and bought.

3) The numbers you have collected are not very statistically significant for either site. Some data is better than none, but it's hard to say at this point what your performance would be going forward.

In terms of people getting to your sign-up form, you may wish to try a service such as CrazyEgg or Clicktale.com CrazyEgg creates heatmaps of websites, whereas clicktale.com actually records the visit, including mouse movements. Both would provide their own separate valuable insights. Clicktale is very expensive, but they have a free version that lets you record 2 pages per visit, which may be enough for you!

Also agreed about the shady SEO stuff at the bottom. That may have saved you a little money, but it has a high risk of tanking any hope of search rankings if Google detects it as manipulating search (which it is). And if a potential customer notices it, it could also be a turn-off.

Best of luck, keep us posted!

FrozenCanuck
03-12-2010, 01:03 AM
I think the economics of your adwords test show you that adwords won't work. You are paying nearly a buck a click, and conversion of 1% seems like it might be typical.

Some things to consider:

1) You'll have repeat customers, so ads have a payback of SEVERAL future orders, not just one. So Adwords might be ok for you.

2) There are LOTS of ways to get traffic aside from Adwords / Adsense / Google. Start following teachers on Twitter and send out links to articles. Post to teacher forums, find marketers who have a list of teachers ... try OFFLINE advertising (magazines), etc.

The test you did shows you that there is actually demand. You already did a lot of the work. You may as well go for it.

Sven
03-12-2010, 06:05 AM
I think that teacher probably do not search for this solution or anything similar but would like to have access to the service. So word of mouth would be important. Add "subscribe to our newsletter" to collect email adresses and "share this site with a friend". I also would want to see some info on pricing before signing up.

Good luck!