Anatomy Of A Stumble

The unprecedented popularity of my recent “Cold Comfort” post was all down to StumbleUpon. For three days, I was in seventh heaven as more and more friendly Stumblers helped my page hits escalate into the thousands — and then along came a reviewer who stopped my party in its tracks.

StumbleUpon logoI’ve written before about the “StumbleUpon Effect”, back in July 2007, when I noticed an increase of several hundred page hits because one of my posts was Stumbled by a handful of Stumblers. Now it’s happened again — in spades. Over the past few days, I’ve watched with mounting excitement as my Cold Comfort post racked up several thousand hits — in fact, by midnight Monday it was around 22,000 since about 5pm Saturday afternoon. This kind of thumbs-up Stumble activity on my little blog is something I’ve never experienced before and left me rather breathless. This time, however — unlike the previous occasion — I’ve also experienced what it feels like to get the thumbs-down treatment too.

StormRegular readers of my blog will know that I’m passionately interested in climate change — the subject dominates my posts. I usually manage around 150 page hits on days when I publish posts. On most other, non-posting, days I languish around the 50 to 60 mark. When I came across a humorous story in Private Eye magazine concerning a man in Austria who was battered on the bum by hailstones shooting out of his toilet — together with mention of freak weather conditions that led the temperature to plunge from 35 degrees centigrade to zero within a few hours, and as a result severe hail storms battered parts of the country, I decided it was material suitable for inclusion in my blog. While being light-hearted, it also highlighted one of the serious aspects of climate change we might all have to cope with in the event of rapid climate cooling: violent storms with accompanying deluges of giant hailstones.

I transcribed the item from the magazine, and while doing so I wondered if anyone else had written about it online. A quick Google search turned up several brief references to the story, including one on another blog that had accompanying pictures which, as Private Eye only had a cartoon, I used to illustrate my own post. I published it on Friday 26th September. Late in the day, Lainiep93 — an online friend of mine — gave it its first Stumble. This then allowed me to Stumble it too (I can’t be the first to Stumble my own posts, as the StumbleUpon system bars me from doing so — but I can Stumble my own posts once someone else has started the ball rolling).

Sunday's StatsThe stats for my web site (provided by GoStats) show that on Thursday, my site had 60 page hits. On Friday — publication and first Stumble day — my site attracted 235 page hits. Saturday’s hit-count was 4,944. On Sunday, things really started popping, with 7,090 page hits being accumulated by the end of the day, and a big increase in hit frequency showing up at around 3pm to 4pm — the reason being that during Saturday and Sunday, more and more Stumblers had given the item the “thumbs-up” (fifty at the current count). These have not been “reviews” as such — only three have taken the time to actually write something about the post, that’s lainiep93, me, and a third reviewer whom I’ll get to in a moment — it was just Stumblers using the “I like it!” button. Once it reached a “critical mass” of Stumblers passing it on to their friends via the thumbs-up and the StumbleUpon system putting it in front of random readers via the Stumble! button, the view-rate went through the roof. There had been 12,034 page hits throughout Saturday and Sunday.

Monday's StatsMonday began as a great day — by 1pm, I’d added a further 7,577 hits, with GoStats estimating 13,966 for the whole day. I was on target for about 26,000 total hits from Saturday up to the end of Monday. I was ecstatic — and then, during the afternoon, my hit-rate slumped, dropping from about 650 in the previous hour to around 100. (You can see the drop in the graph on the right — that sheer brick wall at the 4pm junction.) I believe the reason was empress737700, a Stumbler who gave my post a “thumbs down” and reviewed it thus: GEE, another one that was taken from somewhere else. Jesus Christ people, lets try for some original content, please? This story is bullshit anyway.

The time stamp for Empress’s review is September 29th, 12:31pm. As far as I can ascertain, StumbleUpon’s time stamps use UTC time, the equivalent of GMT — but my stats are logged in BST (i.e. GMT+1 hour), so in my local time this would have been just after 1:30 on Monday afternoon. By the time the StumbleUpon system had processed her thumbs-down and down-rated my post, the effect was being seen in my stats by 4pm BST — 3pm UTC/GMT. (OK — I honestly can’t say whether any other Stumblers had also given it a thumbs-down because, as far as I can see, this info is not provided by StumbleUpon on the review page unless the thumbs-downers also actively review the site in question; so I feel I should qualify what I’ve said by adding that the available evidence points to just this one thumbs-down being the culprit.)

I felt rather deflated. Aside from the fact that Empress had spoiled my stats fun by raining on my parade with her thumbs-down — which obviously had a very negative effect within the StumbleUpon system by drastically lowering the “rating” of my post, reducing the frequency at which it was being randomly placed in front of other Stumblers when they used the “Stumble!” button on their toolbars, and costing me about 3,500 potential hits in the process — she had also completely missed the point of my post. True, it wasn’t original material (how much stuff on the web these days really is original?), but I felt it nonetheless added value to my blog because of the way it highlighted one of the very real consequences of climate change, that of the violent storms it will bring. But of course, Empress didn’t read it in context, just in isolation. She’d obviously come across the same information presented on several other sites, had become bored by the repetition of it — and didn’t believe a word of it anyway. That’s her prerogative — but her single thumbs-down all-but killed my limited opportunity to get my little blog containing what I believe is relevant climate change news in front of a wider audience, so I hope she’ll forgive my chagrin!

Today, it looks like this particular example of the StumbleUpon Effect is wearing off. At this time of writing (mid-afternoon Tuesday), there have been about 670 page hits, with an estimate of around 1,000 for today — still amazingly high for me, but nothing like the heady rush of the past few days, which has seen me achieve more hits in the last 72 hours (22,344) than my blog managed to attract from when I started it in March 2007 to December 2007 (13,581), and more than the whole of 2008 up until a few days ago (13,021). For a little blog like mine, the effect has been nothing short of miraculous and shows what can be achieved via StumbleUpon with the right kind of material — in my case, it would appear that I’m more likely to reach more people with humorous items than I am with my serious posts! It also, however, demonstrates the savage negative effect of a single thumbs-down Stumble. It’s a power that should be wielded wisely — and, perhaps, only after considering the item in question in its context whenever possible.

Yearly Stats

Concerning The Responses My “Cold Comfort” Post Has Received So Far

I’ve mailed Empress asking her to point me to the information that convinces her the original story is bullshit, as I’d be happy to write it up and post it. [UPDATE 03 Oct 08: She's been kind enough to send a reply, which is included in my follow-up post, The Truth Behind 'Cold Comfort'.] Paul Martin commented directly on my post, saying I cry bullshit…..total……idiots (which hardly merited a considered response from me), while qwerty said I’m calling bullshit, the sewer and storm water are separate systems, and why would there be hail stones? why would they be round? (I provided him with an answer.) BenThere then ventured an opinion that This is totally shopped. (Meaning the photos were created using PhotoShop.) I can tell by the pixels and because I have seen many shops in my time. I disputed this, once again asking for more detailed information about what leads him to this conclusion — and so far, surprise surprise, there’s been no reply.

Though I’d be happy to publish anything concrete proving that the basis of the story is untrue, I’m afraid people simply crying “bullshit” and dismissing the story out of hand doesn’t wash with me. I’ll have more to say about this in my next post.

UPDATE 03 Oct 08: Read The Truth Behind ‘Cold Comfort’ for the follow-up post.

The StumbleUpon Effect: Update

StumbleUpon LogoMore concrete evidence showing how StumbleUpon increases traffic to your site. This info is based on events over the past couple of days.

At 6.09pm BST on Thursday 19 July, Daniel Flora read and Stumbled my post titled The StumbleUpon Effect (which I had first Stumbled myself on 22 June). He also reported that he’s seen “a great surge in traffic” to his site labster.net since it was Stumbled. I broke for dinner and returned to my blog at about 10pm. As is my habit, I checked my stats counters and was surprised to see two things: a sudden increase in the number of page views logged by GoStats in the previous three hours, and on the neoWORX map there were about twelve visitors online simultaneously! This was unprecedented activity for my site. Checking the GoStats site for more details, Thursday’s page hit count was already around the 250 mark where it would normally be about 80. Marc Falk had also read my post and commented on it at 9.48pm, saying that he, too, had experienced the same effect on his site after he’d Stumbled one of his own posts — Though I got 2,500 visitors within the first day, it went further the next couple of days. 4 days after I added one of my articles to SU I got 13,000 unique visitors, and my host closed the site because of too much activity … now I got around 500 unique visitors a day. I posted further comments there, including one at 11.12pm, just after the latest GoStats update, saying that I’d observed another 46 page hits in the last hour, made up of about 40 unique visitors … I wonder what I’ll find when I look again in the morning?


GoStats visitor numbers for 20 July 2007

These are the GoStats visitor numbers from just after 8am on Friday 20 July. You can see that in the “yesterday” column, (basically midnight to midnight for Thursday 19 July), it had logged 333 page hits from 239 unique visitors. Already by 8am on Friday morning, a further 445 page views from 335 unique visitors had been logged.


GoStats graph - last 24 hours

Activity had begun picking up at around 5pm GMT Thursday afternoon, which corresponds with the time Daniel Stumbled my post. Activity continued throughout the evening and into the night, hitting a peak of over 100 page hits during the hour between 4am and 5am GMT on Friday morning, after which it began tailing off again. This corresponded with the interest shown in my post by many visitors from the USA, as logged in the GoStats visitor list — it would have been evening time across the US time zones — though there were also many other visitors from across the globe.


GoStats graph - last 7 days

Here you can see the difference in activity between Wednesday and Thursday, with still greater activity logged on the Friday morning (which was only logged up until 8am at that point, remember).


GoStats graph - last 30 days

Here’s the 30-day graph, perhaps the clearest indicator of all that something very unusual happened on Thursday. (The previous spike, around 29 June, was another flurry of activity prompted, I believe, by my Stumbling two more of my own posts, The UK Floods on 26 June, which attracted four comments, and Formula For Arrogance on 29 June, which also received four comments.)

Checking the referrer URLs, almost every single visitor on Thursday and Friday came via the StumbleUpon entry for the post titled The StumbleUpon Effect. (For the record: Friday’s figures ended up at 514 page hits by 381 unique visitors. Saturday’s (so far) have dropped back to my more “normal” values.)

There’s only one conclusion that can be drawn from this hard evidence: Daniel’s Stumble of The StumbleUpon Effect prompted a terrific surge in traffic to my site and it persisted for at least 24 hours before cooling down again.

The Stumbling activity is interesting to follow. Looking at the Stumble page for The StumbleUpon Effect post and picking it apart a little, the initial Stumbles occurred during the first eight days:

Me — Jun 22, 9.42am — 116 ratings, 41 fans
Nightingale2k1 — Jun 22, 10.44am — 983 ratings, 14 fans
Jeffmyuniquelife — Jun 26, 8.05pm — 235 ratings, 22 fans
Incinq — Jun 30, 1.12am — 1,410 ratings, 140 fans

There was then a gap of fifteen days before the next Stumble:

DaBaum — Jul 15, 12.17am — 992 ratings, 25 fans

Then four more days passed before Daniel came along — and from his Stumble came a flurry of activity over 26 hours:

Dflora23 (Daniel) — Jul 19, 5.09pm — 54 ratings, 1 fan
Aurus109 — Jul 19, 7.44pm — 6,830 ratings, 34 fans
Mikemcnary — Jul 20, 2.48am — 3,424 ratings, 5 fans
Lord-xeon — Jul 20, 3.29am — 10,763 ratings, 37 fans
Jiml27 — Jul 20, 3.41am — 3,179 ratings, 9 fans
Altruist64 — Jul 20, 4.46am — 203 ratings, 1 fan
Kancerman — Jul 20, 5.56am — 160,230 ratings, 545 fans
Mattkeegan — Jul 20, 7.32pm — 6,014 ratings, 153 fans

It’s pleasing to see how the ripple spread out from Daniel’s Stumble and seven further Stumblers felt sufficiently engaged by my post to rate it too. However — there’s also something a little disappointing to observe, in that only those seven Stumblers troubled to give the post the “thumbs-up”. Considering there were about 500 visitors who arrived at the post in the past couple of days alone via the post’s Stumble referrer URL, I’d've expected rather more of those 500 to also have Stumbled the post. It appears that they were interested enough to come and read the post (or were presented the post via a random Stumble), but not sufficiently engaged to rate it, despite them having a Stumble toolbar and me having provided a “Stumble It!” button on each post to make it as easy as possible to do so.

It could well be they thought that once they’d read the post, it wasn’t actually good enough in their opinion to earn a ranking. It could equally well be that they were in too much of a hurry to rate it, or were distracted and simply forgot (I’m guilty of this too, when visiting other sites). If only a handful more of those 500+ visitors had Stumbled the post, the activity might have been sustained for longer than the day or so it actually managed.

These observations also serve to underline something that Daniel said in his comment: I think if you have something to offer for your visitors you are much more likely to see your Stumble stats go up. Stumblers are naturally interested in articles about StumbleUpon, especially if written by other Stumblers; blogs like mine, which contain eclectic posts on many different subjects, will most likely achieve only limited success compared to a blog like Daniel’s, which is dedicated to one subject alone (photographs of labradors). It follows from this that, in theory at least, I should be able to maintain a higher flow of traffic to my site by regularly writing (interesting posts) about StumbleUpon! It’ll be fascinating to see if that’s true — the response to this post, providing Stumblers actually Stumble it, will be the judge. Otherwise it’ll probably just be lost in the white noise of the blogosphere.

We all rely on our fellow Stumblers doing their part — my post titled StumbleUpon: Some Answers discusses the limitations StumbleUpon places upon web site owners Stumbling their own pages (unless they create a paid-for StumbleUpon campaign, which not all of us can afford to do). I’d urge all Stumblers to read it — and Stumble it please, because I can’t!

One last thing: a month ago I wrote that StumbleUpon was my third best referrer, with 7.56% of the total referrals. (My own blog main page was top with 50.66%.) Stumble now tops the list with 30.78%, while mine’s dropped to second with 29.79%:


GoStats graph - best referrers

The power of StumbleUpon laid bare, once again, for all to see — and to Stumble!

StumbleUpon: Some Answers

StumbleUpon LogoFurther to my recent StumbleUpon Disaster! post, some answers have emerged courtesy of a couple of Stumble’s help groups.

First, in StumbleUpon Help, where I went this morning intending to post a message outlining my problem, I found that tamburas had already posted a message yesterday with a similar problem: What does it mean when I click “I like it” while looking at a page, the “You’ve Discovered a New Site!” window opens, I fill in the blanks and click on “Submit This Site” and then get a blank window titled “StumbleUpon IE Toolbar”, and the site and my review never show up on my page – ever! Am I doing something wrong?

CH, a “sponsor” and presumably one of the early users (or even a founder) of StumbleUpon, had posted a reply, including this line: We have a limit in place that prevents multiple posting from a single source. Having reviewed your pages, I consider it likely that you are encountering this limit.

I posted a message asking if this could be the same problem I’ve been encountering. CH swiftly replied: If you have more than a handful of posts from the same blog (especially your blog) on the first page or two, then you can reasonably assume that you are affected by the same limitation.

Aha. So that’s probably it, then. I’ve been Stumbling all of my own posts since I discovered StumbleUpon and have obviously reached my multiple posting limit. CH also suggested that the StumbleUpon Group would be a better place to continue my discussion, so I went there too, where I found a similar discussion already going on. You can follow that link and read the full discussion, but basically the StumbleUpon system prevents too many “thumbs-ups” about the same web site or page made by the same Stumbler in order to stop spammers promoting their own web sites too heavily.

I posted a long message in that group containing comments similar to those in my StumbleUpon Disaster! post and received a comprehensive reply from CastorQuinn, another “sponsor”, which answered many of my questions about how StumbleUpon treats main domain URLs and individual post URLs.

So the upshot of all this — and it’s important for all bloggers who are also Stumblers and who want to Stumble their own pages — is, in a word, don’t. If you do, you’ll hit the multiple post limit and, like me, you’ll be wondering why you can’t do it any more.

Instead, we Stumblers must rely on other Stumblers doing the work for us. As CH said in one of his/her replies to me: Use those thumbs wisely, all. It’s *your* community.