June 13, 2008

The Cost of Learning

This one's for Downes... ;-)

Taking a break yesterday from blitzing a whole set of posts to the Visual Gadgets course un-unit (sic) blog experiment, I did a back of an envelope calculation (with a bit of sanity checking from Jon Rosewell and Laura D [hmm - we have new internal blog URLs... "http://www.open.ac.uk/blogs/openair/"]) relating to how much revenue the OU gets "per web page view" for delivering online teaching materials.

That is, if we saw ourselves as web publishers, and we made money from page views/click thrus, what levels of traffic do we need to generate to make the same revenue from 'learning content' pages as we do currently?

Taking the Relevant Knowledge short courses as an example (e.g. TU120 Beyond Google, or T189 Digital Photography), here's an order of magnitude calculation for a 10 week level 1 short course, that takes approximately 10 hours of student time per week to complete.

Assume that the course content pages are what brings the revenue in, not other services (such as assessment, support, etc etc).

Lifetime revenue of each page

Number of 'teaching' web pages per week: 20 Number of weeks per presentation: 10 Number of students per cohort: 200 Number of presentations of course: 10

Number of unique views: 20 x 10 x 200 x 10 = 400,000 (400K) unique views.

Course fee 150 UKP
HEFCE funding  250 UKP (?)
Number of students per cohort: 200
Number of presentations of course: 10

Total Revenue: (150 + 250) UKP x 200 x 10 = 800,000 (800K) UKP

That is, over its life, the page needs to generate of the order 4K UKP (200 x 10 x 2 UKP).

Revenue per student page view

Number of 'teaching' web pages per week: 20
Number of weeks per presentation: 10
Number of unique views per presentation: 200
Course fee 150 UKP
HEFCE funding 250 UKP (?)
Revenue per presentation 400 UKP
Revenue per student page view: 2 UKP (400 UKP / 200)

Annual course revenue - via CPC

Revenue per student page view: 2UKP
Two presentations per year, 200 students per presentation
2 x 200 x 400 UKP =160K UKP per year.
For each page, 2 x 200 x 2 UKP = 800 UKP per year.

Suppose instead we publish 200 'public' pages per year, (say 1 per day for 200 good days per year - imagine, for example, an open content course wiki, which has to be self sustaining, and within which we can manage to post 1 page per day) we need to generate 800 UKP *per page* to make 160K UKP back in that year. For a click thru rate (CTR) on revenue generating clicks of 0.1%, at 0.8 UKP per click (PC) we need 1M impressions *per page* over that year (since 1M x 0.8 UKP x 0.001 = 800 UKP). For CTR of 0.1%, at 0.08 UKP PC we need 10M impressions *per page*...

CPM (cost per thousand) rates

If we were to sell ads on a CPM basis to generate the same revenue that course content pages bring in, the total CPM take per page would be 2K UKP. On a UK newspaper website, the ratecards for ads are of order 15UKP - 30UKP. So we are two orders of magnitude above that (assuming the CPM is for an 'exclusive' ad deal; if there are 10 ads on the page, we are one order of magnitude more costly - but we don't get the volume on the page impressions... (we are at 2 orders of magnitude (?) down on volume).

Timeliness - securing equivalent lifetime revenues per page, within a year

Suppose we want to achieve the same lifetime revenue from a web page in a course that is presented twice a year for five years, within a single year. (That is, we create the same amount of content per uncourse, but need to reclaim an equivalent course lifetime revenue with one year, rather than five. (Why? We assume the content dates over the course of a year..)

If we generate revenue at 2 UKP per click through, and CTR of 0.1%, we need 2M impressions to generate 4K UKP. If we assume 200 effective page viewing days per year, then to generate 4K UKP in 1 year we need 10k impressions per page per day.

If we generate revenue at 0.2 UKP per click through, and CTR of 0.1%, we need 20M impressions to generate 4K UKP. If we assume 200 effective page viewing days per year, then to generate 4K in 1 year we need 100k impressions per page per day.

If we generate revenue at 0.02 UKP per click through, and CTR of 0.1%, we need 200M impressions to generate 4K UKP. If we assume 200 effective page viewing days per year, then to generate 4K UKP in 1 year we need 1M impressions per page per day.

Equivalent CT/CPC Rates

(Thanks to Laura for this one...)

Suppose that rather than course fee/HEFCE income, each course page needed to be self sustaining through click thrus on affiliate links.

To achieve an equivalent page lifetime revenue income of 4K UKP at a CPC of 2 UKP we would need 2000 CTs on affiliate links.

If each page needs to generate 4K from affiliate links, you would need to have every student (cohort size 200) over the 10 presentations click once on an affiliate link on every page of the course.

If the CPC was worth 0.2 UKP, every student would have to click on 10 affiliate links per page to achieve 2K CTs.


Here are a couple more thoughts about the amounts of revenue each page can bring in under different revenue generatng models.

CPM - order of 20 UKP PM, so 0.02UKP PPV
CPC - order of 0.2UKP PC
CPA - (cost per action, cf. revenue per Amazon affiliate link sale) order of 2UKP PA
CPP - "cost per presentatation"(!) - order of 200 UKP
CPL - "cost per lifetime" of course - order of 2K UKP

What's missing in the above blend is a transaction of the order 20 UKP - and maybe this could open up a new business model that balances revenue generation per page with an achievable number of unique content pages and sensible page view volumes?

Blogged with the Flock Browser

Tags: ,

Posted by ajh59 at June 13, 2008 11:24 AM
Comments

A really interesting post Tony - I've done a rambling response at http://tinyurl.com/55u7w5
Martin

Posted by: Martin at June 16, 2008 08:59 AM

Sometimes it’s helpful to temporarily lose the institutional constraints. Have you tried doing something like this to create a tabula rasa? “OU incorporates a new affiliated entity organized as a CIC. Its mission is to provide free learning.” Then start filling in the blanks that describe how that mission might be accomplished.

That’s meant just for illustration. Start with whatever story makes sense. The point is to chew on how you would achieve the mission given no preconceptions. Even if the exercise doesn’t go very far, you may see things in a slightly different way afterwards.

It strikes me that the world could really use some great social innovations. It’s not the technology that’s a limiting condition now. It’s how that technology gets put together. In the Industrial Revolution, the factory was the big social innovation that unleashed the power of the technology. Later, with railroads, the social innovation was the capital markets needed to finance rail development and expansion.

Maybe it’s just a matter of rearranging the existing. Or maybe we need some new organizational forms; or new laws and regulations; or new incentive structures; or all of these and something else as well. I don’t know. But surely it will involve social innovations where the central tenet is the broad public good.

I really hope that advertising is not the social innovation for the web. As a world, surely we can do better than this.

Posted by: Gary Lewis at June 16, 2008 02:36 PM

"Sometimes it’s helpful to temporarily lose the institutional constraints. Have you tried doing something like this to create a tabula rasa? 'OU incorporates a new affiliated entity organized as a CIC. Its mission is to provide free learning.'"

This is partly what we're trying to do with the SocialLearn platform - http://www.open.ac.uk/blogs/socialearn/index.php

Posted by: Tony Hirst at June 16, 2008 03:19 PM