Stats The Way (Aha Aha) I Like It
Better All The Time
By Ged
January 14 2009
Ged Ladd reports on Middlesex Till We Die (MTWD)'s popularity statistics and compares MTWD's growth history with that of the Somerset equivalent on Sportnetwork, Grockles, as well as the recent history of Sportnetwork as a whole. Ged draws some interesting conclusions and makes some canny observations, even though he says so himself in the third person.
In January 2007 I wrote a lengthy piece on Middlesex Till We Die (MTWD)'s popularity statistics to the end of 2006. I shall not repeat all of the statistics I covered in that piece, as you can link to the piece here:
Thanks A Million, MTWD Readers
Where appropriate, I shall update the statistics reported two years ago. More interestingly, I am now better informed as to which statistics matter and I thought it would be fun to compare MTWD's experience with that of our Somerset friends at Grockles and indeed Sportnetwork as a whole. Many thanks to Grockle for providing us with the detailed data for the Grockles site.
Seasonally Adjusted Hit Averages - MTWD
I refer you to the January 2007 piece for an explanation of how and why I use 12 month rolling average monthly hit rates as a key indicator for popularity. Suffice it to say that month on month hits are majorly affected by the time of the season; using a 12 month rolling average adjusts for the seasonal effect. See chart below for MTWD statistics illustrated in simple monthly figures and as a rolling monthly average trend curve:
Figure 1 MTWD Hits Per Month *
At the end of each calendar year (for which the rolling average monthly hits statistic is available) that figure was:
December 2003 - 24,861
December 2004 - 37,807
December 2005 - 52,475
December 2006 - 92,642
December 2007 - 155,939
December 2008 - 225,173
I named the January 2007 piece "Thanks a Million". It took MTWD just over three years from its inception (May 2002) until June 2005 to achieve its millionth hit. (Had we been able to trap that hit, we'd have no doubt awarded a miniscule prize for it). It took from May 2002 until October 2006 to break the rolling annual hits milestone of a million hits per annum. Currently (using December 2008 figures) that figure is already 2.7M - in other words at current rates MTWD gets a million hits every four-and-a-half months or so.
This represents considerable growth.
MTWD and Grockles Compared
In May 2002, when life started on Sportnetwork, Grockles was a somewhat more mature site than MTWD, attracting thousands rather than hundreds of hits per month. The following chart repeats Figure 1, overlaying Grockles stats for the same period:
Figure 2 - MTWD and Grockles Compared *
Grockle sometimes tells me that he thinks Grockles has reached a popularity plateau whereas MTWD is still growing. This chart indicates that Grockle is mistaken in this view. While Grockles did go through a brief period of slow-down (December 2006 to June 2007) the general trend is for Grockles to have grown in similar fashion to MTWD, more or less always keeping slightly ahead.
Regarding the 2006/7 slow down for Grockles, I think it coincides with a period of technical difficulties for Sportnetwork and the port to the new Sportnetwork platform. Perhaps MTWD readers tended to be a little more tenacious (or, let's not mince our words, potty) at that time. As you will see from the final chart, even MTWD's rate of growth dipped a little at that time.
In any case, you look at the way the rolling average hit rates have been tending in the last 18 months, and you would say that Grockles has plenty to celebrate.
MTWD, Grockles and Sportnetwork Growth Rates Compared
Our final chart illustrates comparative rolling average hits growth rates for MTWD, Grockles and for Sportnetwork as a whole. The Sportnetwork figures (which include MTWD and Grockles of course) cover all of their sports sites, including football, rugby and motor racing as well as cricket. MTWD and Grockles combined comprises some 2% to 3% of Sportnetwork activity. Given the size disparity, it seemed to me that rolling annual growth rate percentages was the most useful statistic to use, as it eliminates seasonal effects and is not absolute size dependant.
Figure 3 MTWD, Grockles and Sportnetwork Growth Compared *
A few observations on the above statistics:
- Grockles rolling average growth rate is considerably more volatile than either MTWD or Sportnetwork as a whole. If that is a plateau, it is a pretty hilly one!;
- For most of MTWD's early life, its percentage rate of growth was similar to that of Sportnetwork; since mid 2006, however, Sportnetwork's rate of growth has been tailing off whereas MTWD's rate of growth has remained strong, even during that period of technical difficulties;
- Sportnetwork went into "popularity recession" for a while in 2008, although there are signs towards the end of 2008 that it is starting to recover.
A Few Further Thoughts and Questions for the Audience
As usual, if any of the professional statos would like to comment on the meaning of these stats, those comments would be much appreciated. Or indeed any other amateur statos with a point to make. Please post your thoughts on the thread or e-mail us.
If anyone wants a look at the underlying data, it is available at:
http://spreadsheets.google.com/pub?key=pwWi9CpG2dshENNfUUHC2WQ
I have made this point many times before but it is worth repeating here: MTWD and Grockles are not in competition with one another. I am convinced that either site would be much the weaker in the absence of the other. In particular, during the off-season (which is most of the year for cricket), both sites benefit enormously from the cross-border traffic and cross-fertilisation of ideas.
My final point is that popularity statistics are not the be-all and end-all of our sites. To use an unpleasant metaphor, many people rubberneck when there is a road crash on the other side of the road, but that does not make road crashes a good thing. Similarly, although flames and rows might cause a short-term surge in hits for discussion group websites, I am convinced that, in the medium to long term, orderly debate (which can include banter but not harassment and bullying) is necessary if sites such as ours are to thrive.
Many thanks to all of you, because it is the readers and posters who ultimately determine the success or otherwise of our sites. And judging by these stats, both MTWD and Grockles can categorically describe themselves as a success; at least for now!
* Many apologies for the poor picture quality. This probably says more about my ability to manipulate images within the constraints of the Sportnetwork editing system than it does about that system. here is a link to a Googledoc of the presentation quality copies of the figures:
http://docs.google.com/Presentation?id=dgnf6nkc_103dv4rk6gt
Enjoy.
View a Printer Friendly version of this Story.
Digg
Delicious
Facebook
Reddit
StumbleuponIf you do not already have an account Click here to Register.
[unstuck by admin]
Edited 1 time(s). Last edit at 2009:01:23:08:49:05 by Ged.
I'll fiddle around with the pictures later. In the meanwhile, apologies and please go with the flow!!
I often came over to read your site and now I post as well, I will not say the reason for this.
It is good to read the different views on our great sport and congratulate Ged and the others for running a great site.
And now I will head back to Grockle or he will miss me.
.
The link to my Photos on Photobucket - [s1152.photobucket.com]I entirely agree with Ged's comment that "orderly debate (which can include banter but not harassment and bullying) is necessary if sites such as ours are to thrive". In my view, the moderators, on this and the Grockles sites, are the (often unsung) heroes in helping to set the right tone. In my view, that is the key reason why these two sites are - in my view - infinitely preferable to other forums I have stumbled across.
I visit to be entertained and amused as well as to be informed. I would not bother logging on - let alone contributing - unless discussions were conducted in a civilised manner. Long may this continue on both sites.
The one thing thats always amazes me is the number of people who read related to the number of people who post. It seems so easy but the percentage that post is a very small percentage of the general readership.
However one needs to explain that'hit rates' does not relate to a number of people logging on but a count of every page a person reads. We have a very 'sticky' bunch of stalwarts they tend to log on and then stay about browsing through stuff.
The number of people logging on is a hard one to measure. I got quite hung up on "number of IP addresses" when I did the analysis in January 2007, but have since been informed by Sportnetwork and others that the stat is unreliable.
Technology changes (e.g. roaming IP addressing) and demographic changes (e.g. access to internet through multi-IP sites such as public libraries and internet cafes) render the IP addresses stat nigh-on meaningless as a trend stat.
So hit rates it has to be.
And as we have discussed elsewhere, hit rates do not necessarily have a strong correlation with number of postings. Quality of postings affects hit rates as much if not more than quantity of postings. That's because, as you say, Grockle, the majority of readers are lurkers or just occasional posters.
My current rough estimate for MTWD is as follows:
- a few dozen addicts
- a few hundred regulars
- a few thousand unique visitors per annum.
I suspect that Grockles is similar.
A few dozen 'unique' addicts perhaps?
Pander to your masses ...
Yours, hurt of Somerset.
VtF.
Edited 1 time(s). Last edit at 2009:01:15:01:57:39 by bloke.
Please now all say in Unison, "yes, we are all individuals".
Thank you.
Ah crud !!
Wrong thread !!!
Perhaps I should hit each thread I read three times!!
Perhaps I should hit each thread I read three times!!!
That would up the hit rates.
Hippity is author of top features: Hippity's Tale and Bunny Away
