Recent Blog Entries
On Traffic, Money, And Advertising
Written by Mike Rundle on August 24, 2006
I was talking with my friend Pat McCarthy from Right Media the other day about online advertising, and one of the things we chatted about was that one of the best-known “Web 2.0″ companies out there, Digg, is having trouble selling ads. About 30 minutes after I got done talking to Pat, I went to Digg.com and for the first time in my entire life, I saw a banner ad at the top that wasn’t a Google ad. Gasp! They sold one!
According to their advertising profile on Federated Media’s website, they have (approximately) 5M uniques a month, which is an extrapolated number that corresponds to about 15-20M pageviews per month. I don’t know how old those numbers are, but let’s work with them for a second. Their rates for the top banner are $16CPM, or cost per thousand impressions, so to launch a campaign for just one day on Digg.com it would cost about $10,600. One week is $74,000, one month is $318,000. I think the reason that few companies advertise on Digg is that the companies who can actually afford to advertise there would rather advertise on Yahoo!, CNet, ZDNet, NYTimes, LATImes, etc. The only companies that can afford to run ads on Digg are those with gigantic advertising budgets, and the companies with gigantic advertising budgets go for more mainstream sites that have a more proven return on investment.
So if the poster child for “Web 2.0″ has trouble selling advertising, and “Web 2.0″ is all about giving services away for free and selling ads (tongue in cheek), what hope is there for the rest of us?
Seth Godin’s Alexaholic List
Marketing genius Seth Godin recently hooked up with traffic rank site Alexaholic to produce his Web 2.0 Traffic Watch List, which is ~1000 companies and websites all ranked according to their Alexa ranking, a good indication of their total amount of traffic. Alexa rank is not a perfect science because only those with Alexa software installed can impact the data, but it does provide a real-time look at traffic trends and is useful for comparison research.
On the gigantic list of “Web 2.0″ companies, little old 9rules ranked #131 and beat out notable companies like Rollyo, Prosper, Riya, Zimbra, Consumating, and even Microsoft’s Start.com. Although we don’t get much love from the West coast, we have been gaining traffic rapidly this summer which is always nice to see:
b5media, Leading The Pack?
We don’t really put ourselves in the blog network rat race anymore, but our friends at b5media have recently begun publishing their aggregate network traffic numbers at the tune of 20M pageviews per month, a figure that definitely needs to be analyzed.
To give some perspective, 20M pageviews per month is more than the total pageviews received by TechCrunch (3M), GigaOM (~750k-1M), Boing Boing (~9M), and Newsvine (~800k-1M) combined, with a few million pageviews to spare.
B5media has about 150 weblogs across various content channels, so if you do some extrapolated math, you’ll find that (on average) every one of their weblogs must receive 133k pageviews per month equaling about 4,400 pageviews per day. Four thousand pageviews per day for every single blog is an awful lot, especially considering a lot of their blogs go for the niche audiences like Gryffindor Gazette or Flu Patrol, so either a ton of their weblogs get even traffic or they have a few homerun blogs that get the bulk of their 20M pageviews per month. I think the latter scenario is more plausible, so let’s find some of these homeruns.
Blog Network List shows the Technorati rank of all their blogs, so if we’re going with the “a few homeruns” theory then some of their sites need to have some pretty high Technorati juice in order to pull down enough pageviews for their 20M per month. After checking out the list they only have 10 blogs that crack the Technorati 10,000, but 3 out of those 10 include ProBlogger, Ensight.org, and DuncanRiley.com, which are the three sites owned by the founders. Out of the 10 in the Technorati 10,000 that aren’t owned by the founders (or aren’t their company site b5media.com) we’re left with The Gadget Blog, Successful Blog, Genetics and Health, Emerging Earth, eBeautyDaily, and Cellphone9 as their top blogs in the entire network.
Unfortunately, Technorati rank doesn’t mean much when you’re talking pure traffic, so for a comparison we’ll have to move to Alexa. As a baseline, I’m going to compare all these top b5media blogs to my lowly BusinessLogs.com site which only receives 1,000-1,200 pageviews per day. Here’s BusinessLogs vs. the first 4 and the final 2. As you can see, the only blog that comes close to the 1k/day Business Logs traffic is Successful-Blog, but that makes sense because its Technorati rank is currently better than Business Logs. What is odd to me is that out of the 150 blogs b5media runs, the ones at the very top (that should be responsible for the bulk of their 20M pageview per month traffic) can’t even breach the Business Logs Alexa results and I only get 1k pageviews per day. Even if you add in the founder’s blogs and give them healthy traffic estimates (let’s say Ensight, ProBlogger, and DuncanRiley all get 20k pageviews per day) that only tacks on 1.8M pageviews per month to their bottom line. 20M pageviews for the entire network minus a 1.8M pageview estimate for the founder’s blogs, minus a healthy 1M pageview estimate for b5media.com itself means that the remaining b5media blogs run by their editors receive ~17.2M pageviews per month even though out of all those remaining blogs, only 6 make it into the Technorati 10,000. Yes the data at Blog Network List might be out of date, and yes Alexa stats aren’t a perfectly reliable source, but these small variances seem underscored by the large mathematical gaps I’ve just presented. Of course I’m probably an idiot and am missing something big here, so Jeremy, please show me the light by linking to some screenshots!
Social Networking Gets You The Traffic
The problem with “Web 2.0″ (well, one of many problems) is that many companies are generating solutions and companies around those solutions, even though the problem they’re solving hasn’t been discovered yet. Too many startups in the Valley are concentrating on getting the core tech audience but forget about the 99% of the world who don’t care about Ajax or your API. The startups doing the best happen to be the social networks, namely, MySpace and Facebook, because they’re appealing to the majority audience instead of just the tech-savvy few. You don’t need to know about RSS or Ruby on Rails to appreciate what MySpace and Facebook are doing, and that’s what brings the masses to their front doorstep.
These two social networks have just inked deals with Google and Microsoft respectively that will give these two advertising giants the exclusive right to sell advertising on two of the largest social network sites around. What’s interesting about this is that instead of having Digg’s problem of not being able to sell ads, these two social networks were negotiating with major companies to sell all their ads at once! Why the complete polar reversal? Social networks attract people of all tech skill levels, from all aspects of the socioeconomic ladder, and compared to Digg they have an extremely mainstream audience. What gigantic blogs do the best? The ones that attract a mainstream audience, like pop culture blogs, gossip blogs, etc. The key to growing out of your niche is to provide value to an audience outside of your core, and the companies/websites that can do that always seem to be the most successful.
August 24th, 2006 at 10:58 am
So faulty base info (Boing Boing gets well more than 9M pages per month, per comScore ratings) + a rating system that isn’t tied to traffic (Technorati) + a traffic system that isn’t accurate (Alexa) leads you to say we’re off the mark? Oh, and BNL is more than a month out of date.
I could break it down for you, but there’s no reason to do so. The data’s accurate.
August 24th, 2006 at 11:08 am
Seriously man, break it down. You can’t talk numbers without being able to back them up. Denton talks numbers and has SiteMeter on every blog, Calacanis (before WIN was purchased) had graphs next to numbers. 20M pageviews per month on your network is a gigantic number, take Boing Boing out (FM stats page is very wrong for them I now see) and you still have more traffic than the largest tech blogs in Silicon Valley and many startups, combined. You can’t just drop a number and say the data’s accurate because you say it is, I think you owe it to your readers and editors to show just how large b5 looms in the eyes of visitors.
Technorati rankings are valid when we’re talking blog traffic — show me blogs ranked ~10,000-50,000 that receive hundreds of thousands of pageviews per month and I’ll quiet down. My singular question is related to traffic and links, if a blog has a gigantic amount of traffic but not many links, is that possible?
August 24th, 2006 at 11:17 am
It’s possible for a site to have a lot of traffic and not many links if the traffic is mainstream, but I would think in that case there would be a strong Alexa rating. Bloggers for the most part don’t use Alexa and Alexa is built into IE6, which has the largest browser market share.
As an example, compared to Business Logs ST has less reach but has more page views. That Alexa perception is dead on.
August 24th, 2006 at 11:20 am
Mike, Technorati links aren’t even accurate for traditional blogs: http://avc.blogs.com/a_vc/2005/08/counting_links.html.
How do you expect it to be accurate when most of the traffic on some blogs is from forums and not from blogs?
Links != traffic. And if Technorati’s links are off by more than 50% (and they are) it’s even worse.
We don’t owe it to readers to explain our traffic, since our traffic doesn’t impact readers. We owe it to advertisers, which we do. We owe it to partners, which we do.
But, if all it really takes is one screenshot from a high traffic blog that isn’t in the Technorati Top 100, then here you go, from PittWatch (for the record, it’s 9,883 on Technorati now but only 11,194 on BNL):
http://www.ensight.org/uploads/9rules.jpg
August 24th, 2006 at 11:26 am
“isn’t in the Technorati Top 100″ should read 1000, per your request. Yes, it’s possible. Technorati’s inaccurate, BNL is inaccurate, FM is inaccurate and Alexa is inaccurate. You can’t use 4-5 services each off b5 30-100% and hope to arrive at a picture of our traffic. Nor should you. It shouldn’t matter to anyone but advertisers what our traffic is, since it doesn’t impact anyone but advertisers
August 24th, 2006 at 11:27 am
Grr, “each off b5″ should be each off “by”. I hate new keyboards
August 24th, 2006 at 11:34 am
Jeremy, if your stats are of no interest to anyone but your advertisers, then why continue to talk about them on your blogs? I’d think that out of anyone connected with b5, you owe it to your editors most of all to prove statistics because they’re paid based on advertising sold from the traffic.
I can’t tell a lot from that screenshot, but if PittWatch does get 400k pageviews per month then that’s about what I expected for dozens of your blogs to have. Like I said, 20M pageviews across the network averages down to every single blog receiving about ~130k pageviews per month, so what about the other 140+? I’m pretty sure that you have a few blogs leading the pack in terms of traffic, but that doesn’t add up to 20M.
August 24th, 2006 at 11:38 am
Mike, you asked for a blog in the top 1000 with hundreds of thousands of pages. PittWatch isn’t the biggest blog. And, our editors have access to network stats, so we don’t need to explain anything to them.
I’ve proven all I need to prove, since we’ve only talked stats once (that I remember) in the last few months. It’s just not something we talk about, though we’ll likely celebrate when we hit some major milestone like 50M pages/month or our 1Mth comment or something.
I have no motivation for sharing any more than this. Everyone who needs to know does. End of story. Feel free to enjoy speculating, though.
August 24th, 2006 at 11:39 am
The problem with catering to the mainstream if your niche does not naturally fall into that category is to risk dumbing down your content to the point that your core audience no longer has any interest. Rather than catering to the masses, why not concentrate on putting together a good product? A List Apart, a site that easily fits into the realm of a ‘niche’, generates a tremendous amount of traffic - more than most ‘mainstream’ blogs its size. One reason, in my opinion, is their high standards for submissions and overall top-notch quality of presentation and content.
The other problem with entering the mainstream is that you are now competing with both larger sites and larger group of sites. The audience of such sites seem to be much more finicky as well since there are a plethora of sites they can go to.
If one’s site naturally fits in the mainstream, good for them, but to force yourself into the mainstream seems dangerous at best and you may just end up replacing savvy loyal users with half-hearted uncaring ones.
August 24th, 2006 at 12:05 pm
Sorry man, we’ll have to agree to disagree. The simplest way of proving me wrong (or proving anyone wrong) is to show that they’re wrong and unfortunately you’re not doing that. Just in the past few months I’ve seen references to your traffic figures here, here, here and other places, but I’ve never once seen a graph or chart to accompany a traffic figure.
AdBrite sales for the entire b5media network only add up to 2M pageviews per month (these are not run on every b5 blog, but I randomly looked at 25 and only saw 1 without AdBrite), and then BlogAds traffic figures for PittWatch say 17k pageviews per week, which is a far cry from 400k a month. If these stats only matter to your advertisers, then why do your advertisers have drastically different numbers than you’ve presented?
August 24th, 2006 at 12:14 pm
Again, no motivation to Mike. You can try and pick a fight all you want. I’ve shown that Technorati and Alexa are so far off the mark it’s not even funny. You can dig all you want, but unless you have server stats, or actual ad impression stats, you are just making noise. It’s entertaining noise, but it’s just noise.
Why don’t you go ask Boing Boing why they have 9M pages/month in FM but they claim roughly that many uniques? When you can answer that question, then we’ll talk.
August 24th, 2006 at 12:23 pm
I can answer the question right now, at the bottom of FMs profile (and every site’s profile) they say that they cannot accurately track uniques so they extrapolate it. The figures for Boing Boing say “more than 2.125M uniques” which is an extrapolated number based on their monstrous pageview count.
Jeremy, you keep talking about how your traffic is for your advertisers only, so the motivation to explain why you say PittWatch gets 400,000 pageviews per month when your advertisers count it only getting 72,000 should probably be there. How are your server stats relavant when your own advertisers see that they’re off by a factor of more than 5? That’s what I’m getting at here.
August 24th, 2006 at 12:24 pm
FM is approximating and tries to underestimate ad impressions whereas your stats are inflating them by 5x.
August 24th, 2006 at 12:31 pm
And Jeremy just to be clear, I’m not picking a fight with you, I’m doing this so readers are getting the facts. I’ve done this at Business Logs many times because I just try to call it like I see it.
August 24th, 2006 at 12:39 pm
If you want facts, ask for them. If you’d emailed or called and said “this doesn’t add up, can I see your stats just to ensure I’m not insane”, I might have just said yes. But publicly arguing isn’t the same thing.
Not like you don’t have my email, phone and Skype.
It would take me 5 minutes to show you and explain it to you, without revealing any terribly sensitive info.
August 24th, 2006 at 12:39 pm
jeremy if you feel they are so wrong then how can you possibly explain the pittwatch blogads stats?
advertisers look to those stats to get an accurate picture of the site’s traffic, and where are you getting this 400k figure from? your server stats are totally bunk, you probably use awstats or some other inflating app.
yeah awstats says my blog with 200 pageviews per day gets 5k so shit, i’m a millionaire!!
August 24th, 2006 at 12:41 pm
And Mike just to be clear, if you received proof of traffic, you’d publicly apologize right? Since you’d gotten the facts and all, right?
August 24th, 2006 at 12:45 pm
Lol I put this in a public forum because you’ve talked about your traffic in public forums, it only makes sense to keep it a public conversation. I think it’s fair to everybody to see some pushback.
If you could show me how BlogAds (your advertiser) sees PittWatch as having 5x fewer pageviews per month than your server stats say, or how your AdBrite stats are much lower (not on all blogs, I realize, but still far too low), then I’ll absolutely apologize publicly.
August 24th, 2006 at 12:58 pm
I do agree that unless there is access to server stats, it’s hard to tell (and those stats have to be properly analyzed but Jeremy said in our chat that he has taken measures to help make the stats more valid). 3rd party stats can’t be 100% accurate but they are usually (when all factors are considered) a good indication. There are some large discrepancies here, that perhaps there is a logical explanation. Because the stats were spoken about publicly I don’t think it’s picking a fight to question them publicly. It’s being questioned in the manner it was originally presented. This could be a learning experience for everyone involved.
The lesson here is if you state your stats publicly then you’ll most likely be questioned about it publicly.
August 24th, 2006 at 1:15 pm
There have been thousands of pages of blog posts written in the last month about differences even between comScore and Nielsen. Differences of 50-80%.
Publicly stated doesn’t mean you get to see every bit of sensitive information. And people not knowing how different stats packages calculate stats doesn’t mean I have to explain it. If you’re unsure why Webalizer is 10x what Analytics is, or why AWStats is 1.5x Webalizer, or why PhpAdsNew is 0.8x Analytics or why AdSense is 1.2x PhpAdsNew, that isn’t my problem.
I’ll give you an example though, in the hopes of helping your readers understand a bit of the challenge with “stats” and why third party stats are so important (wrong ratios, but ah well).
ENSIGHT
Webalizer: 218K pages so far this month
AWStats: 223K pages this month
Analytics: 83K pages this month
PhpAdsNew: 24K pages this month
AdSense: 59K pages this month
Oh, and it does about the same traffic as BusinessLogs according to Alexa, which, according to Mike, means it should get between 30-60K pages/month.
So when you drill down, you find that more than 30% of all requests are from feed readers. 20% are from search engine bots. The rest, between 100-150K pages, are pages that are actually served to a browser (give or take 20%).
So, figure on 100K, even, pages being served to browsers. Now, some will have ad blockers on (notoriously bad for blocking AdSense and PhpAdsNew) (10-40%, depending on the audience). Others will have JavaScript turned off (5-7%).
What does that leave us with? Well, a spread of between 50K-200K/month in terms of actual pages served.
And that’s just the discrepancy on 1 blog.
Oh, and on Blogads it only registers 23K impressions and on Adbrite it’s less than that. Why? Well I don’t know, since they don’t use industry standard, publicly available algorithms. But, considering how many times I’ve personally seen them not load on our blogs maybe that’s why. Or maybe not. Who knows, since it’s not public?
Either way, a factor of 5, 10 and even 20 is totally normal between server stats and actual impressions. It’s even worse if people don’t have their server stats to differentiate between pages requested and pages served. In that case it can be a difference of up to 50x.
So do I need to “prove” anything? No. Because even if I did, there’d be no way to certify the results as accurate. I could show AdSense, Analytics, AWStats, Webalizer, PhpAdsNew and still not convince you.
However, the lowest stat package we have available shows about 10M impressions so far this month. The biggest number puts it at more like 35M.
Which is why we don’t normally talk about our stats. The only reason we did in the call for business bloggers is that 5 of them asked, so we lowballed a generic number.
To be clear, if we commissioned a comScore study, they’d probably say something around 7.5M, but it could be as high as 40M (based on my work with them in the past). So it wouldn’t even help you out then. You might believe it more, understandably, but that wouldn’t make it more “accurate”.
Accurate stats are like security on Windows. You never really get it, but even if you thought you had a handle on it there’d probably still be something wrong.
Which is why stats just don’t matter. Consider how often JCal talks about his stats (weekly, give or take). Consider how often Nick does (monthly, give or take). Then consider how often we do. We just don’t care. As long as we deliver the impressions for advertisers (using industry standard algorithms), nobody else should either.
And that quick lesson (there have been books written about this subject, and week-long seminars I’ve attended) is how I’ll end this.
If you really, really feel like poking at a number which we only said because pressured to, and really don’t care about, knock yourselves out.
August 24th, 2006 at 1:21 pm
For more info, even amongst the highest traffic sites, with third party validation, here’s some reading:
http://webmeasurement.wordpress.com/2006/08/10/siliconbeat-beats-webanalytics-industry/
http://www.siliconbeat.com/entries/2006/08/10/web_stats_are_broken_so_youd_better_have_brass_knuckles.html
http://washington.bizjournals.com/washington/stories/2001/04/09/newscolumn10.html
Effectively, comScore gives you the accuracy of Nielsen on TV. Which is to say that nobody knows for sure how much each person watches, but Nielsen’s as good as it gets. Ditto comScore, in spite of the spread.
August 24th, 2006 at 1:28 pm
Well Jeremy, in the podcast you’ve said that you’ve filtered out bots, feedreaders, etc. out of your own stat programs, so where does that fit in? I’d say that the most accurate results would be coming from your own advertisers on PittWatch (just refreshed 30x, BlogAds on every view) and they’re 5x lower than what you’re talking about publicly on your blog, but you make no mention of that when you talk about it…. only on this blog when I call the numbers out.
“Either way, a factor of 5, 10 and even 20 is totally normal between server stats and actual impressions. It’s even worse if people don’t have their server stats to differentiate between pages requested and pages served. In that case it can be a difference of up to 50x.”
Okay, so when you say you have 20M pageviews across the network does that mean you have 4M actual pageviews, actual pageviews being somebody going to a page and actually loading it? If so, then how can you publicly state a number that you know could be inflated over 5x? This is precisely what I’m finding when I look at BlogAds or AdBrite stats for your network — so shouldn’t you be reporting those numbers and not the inflated versions?
The question I’ve been trying to ask is how you can feel comfortable saying 20M pageviews when your own advertisers (the only ones who matter, right?) count their ad impressions on your pages and they are significantly lower. Significantly lower, both BlogAds and Adbrite. That’s the question. What point is there in telling your advertisers you have 20M pageviews when by their own count you have 5x less??
August 24th, 2006 at 1:34 pm
I get the feeling you’re trying to dodge my question, so I’ll restate it as clearly as possible:
If your advertisers count 5x fewer pageviews than your own figures, why even mention your own figures if (by your own account) server stats are flawed? Why not just go by the figures they see since that’s all they care about?
August 24th, 2006 at 1:46 pm
1. Show me a blog with low links and high traffic and I’ll back off. Done.
2. Explain to me how there can be a major discrepancy in traffic and I’ll back off. Done.
August 24th, 2006 at 1:54 pm
Jeremy, you need to go into politics because you cannot answer a simple question. If your traffic figures are only there for your advertisers, and your own advertisers count PittWatch’s traffic as 5-6x lower than your server stats, how can you say in good conscience that they’re wrong. How? It makes no sense.
You can show me all the sites you want about how server stats are misleading, or how they’re bloated, but it just goes on to prove that your own figures are inflated. Your advertisers say you get 5x less traffic than you like to tout, and that’s the bottomline. Who gives a shit if your server stats say X when your advertisers say you get 1/5th X. That’s like me saying I’m a scratch golfer but my scorecard continually comes up as me not breaking 100.
August 24th, 2006 at 1:56 pm
One last time, because I do generally respect you. I took the time to explain an industry to you in 1000 words and all you can do is nitpick on the stupidest things.
So, which stats should a CEO talk about Mike? The ones that are industry standard and accepted as having a reasonable degree of accuracy, or the ones everyone scoffs at and the founders themselves don’t trust?
1. BlogAds: inaccurate, non industry standard algorithms, no respect in the stats industry, low pages
2. Adbrite: inaccurate, non industry standard algorithms, no respect in the stats industry, low pages
3. Google Analytics: inaccurate, but a nice algo, industry verified, reasonable respect in the stats industry, lower pages
4. PhpAdsNew: inaccurate, nice algo, verified algo, reasonably respected in the stats industry, lower pages
5. Webalizer: inaccurate, nice algo, stats industry respect, higher pages
6. AWStats: inaccurate, okay algo, industry standard, verified, highest pages
Should I really, really say stats that I know aren’t accurate, aren’t industry standard, aren’t verified?
No. As a business owner, I say the ones that are as close to “the truth” as possible. Saying those are the wrong ones just because an 1) inaccurate, 2) unrespected, 3) no history in the stats industry, 4) unverified algorithm, 5) private algorithm has lower numbers?
That would be irresponsible.
Posting the stats that are as accurate as possible without paying the massive comScore fee? That’s responsible Mike. Becuase I go with where industry trust is. Ad industry, and stats industry. And both have more trust in Analytics and Webalizer than BlogAds. So, when we talk stats, those are the types of numbers we’re pulling.
Again, if you have a problem with us going with the industry standard packages that isn’t my problem. There’s no dodging, but until you get over the fact that the stats you’re seeing aren’t accurate there isn’t anything I can do. When you realize they’re inaccurate, maybe you’ll be ready to look at the logs from one of the servers.
Until then, this all just wasted breath. I spent more than 1000 words educating you on the stats industry and you didn’t even respond to that. I didn’t specifically address Blogads and Adbrite because they aren’t accurate.
August 24th, 2006 at 2:01 pm
Jeremy, you say your traffic is only relevant to your advertisers, so do they really care about what stat packages you use? No, they look at BlogAds or AdBrite and see exactly how much traffic they are paying for when they buy a link or a banner.
If BlogAds and Adbrite are so innacurate, why use them as advertisers across your entire network? If they’re telling companies that you only have 1/5th the traffic that you say you get, wouldn’t that piss you off? Cause you to drop them in favor of something else? Aren’t they losing you money if your actual traffic is 5x higher than what they say?
August 24th, 2006 at 2:03 pm
Mike: Google has more than 7x the Blogads or Adbrite stats for the network. Again, who should I believe? Google, who has a history in stats (having bought one of the most respected stats companies in the world), or Blogads?
Is that what you’re honestly, honestly saying?
Your golf analogy’s a good one.
Blogads: You shot a 12
Me: It was an 18 hole course, I couldn’t have shot a 12! I shot a 62.
My scorecard: 68
Google (ie: official judge at the tournament): 70
Which score would you say? I’d say the 70.
You ultimately have 3 issues here Mike:
1. You think I’m lying to you
2. You don’t trust the stats industry that Google is better than Blogads
3. You are trying to pick a fight
Let me ask you, when 9sponsors sells ads, do you quote your server stats, your PhpAdsNew stats or some third party stat that you know is wrong?
Because that’s what you’re asking me to do. To pick every stat package that is even vaguely proven as accurate and choose one that you can prove is absolutely, unequivocally wrong.
August 24th, 2006 at 2:05 pm
“If BlogAds and Adbrite are so innacurate, why use them as advertisers across your entire network? If they’re telling companies that you only have 1/5th the traffic that you say you get, wouldn’t that piss you off? Cause you to drop them in favor of something else? Aren’t they losing you money if your actual traffic is 5x higher than what they say?”
Because our advertisers get access to actual impressions actually served by the ad server. We’re dropping Adbrite and Blogads. Neither pays based on traffic. In both cases it’s a term payment, so the actual number of ads doesn’t matter.
Why you continue to insist these are accurate is beyond me. Inaccurate. Inaccurate. Inaccurate. You hear that AOL cancellation phone call? I feel like that guy. Inaccurate. Inaccurate. Inaccurate.
August 24th, 2006 at 2:12 pm
“Webalizer: 218K pages so far this month
AWStats: 223K pages this month
Analytics: 83K pages this month
PhpAdsNew: 24K pages this month
AdSense: 59K pages this month”
Well for Ensight.org you have Google (analytics and adsense) coming in well under AWStats, and that’s my point. You screenshotted me PittWatch’s traffic and that came from AWStats, which as you’ve just said, does not track traffic as accurately as Google. So when you gave everyone the 20M figure for all of b5media, was that from AWStats as well?
August 24th, 2006 at 2:12 pm
“So when you gave everyone the 20M figure for all of b5media, was that from AWStats as well? ”
No, because that would be irresponsible.
August 24th, 2006 at 2:18 pm
Where’d it come from? Was that Google Analytics? Why would you try to prove your point against me by screenshotting stats you know are inflated?
“You ultimately have 3 issues here Mike:”
Haha nah I’m much more simple. My one issue is that I just don’t believe you and I’ve found satisfactory, third-party evidence that is on my side. You’ve yet to show me how you have 20M actually loaded pageviews per month across the network, and until then, I’ll continue to not believe you.
August 24th, 2006 at 2:19 pm
And yes, we use PhpAdsNews stats because we’d rather underestimate ad impressions (FM’s approach) then possibly give the wrong picture by using AWStats or Webalizer.
August 24th, 2006 at 2:21 pm
“satisfactory, third-party evidence that is on my side”
Inaccurate. Inaccurate. Inaccurate. And there’s your problem. Because you haven’t. Inaccurate. Inaccurate. Inaccurate. Every bit of third part evidence you’ve chosen is inaccurate.
August 24th, 2006 at 2:24 pm
Well Jeremy if I can’t believe what your own ad brokers are telling potential advertisers then I guess there’s no point to traffic stats in general.
What service did you use for your 20M statistic? Google Analytics or AdSense, the two most accurate sources you say? Screenshot me. That’s the only way you can prove it’s not inflated. That, or put SiteMeter on all your sites and make your stats public, but making inflated stats deflate wouldn’t be any fun I’m sure.
August 24th, 2006 at 2:38 pm
Oooh, look at all the attitude. It’s amazing how well you recover after being wrong half a dozen times in one conversation.
You said *you* trust your PAN stats. So here are our PAN stats.
http://www.ensight.org/uploads/9rules2.jpg
So, after your third promise to back off, will you now?
August 24th, 2006 at 2:42 pm
“I guess there’s no point to traffic stats in general.”
Bingo!
Only took 20 comments and 5000 words to get you to realize that stats don’t matter. Everything’s relative. As long as you *do* serve the ads advertisers buy, though, everything’s good. Beyond that, nobody cares.
August 24th, 2006 at 2:46 pm
Yup, done now. Our PAN stats look like that as well with multiple ads per page. Nobody cares about stats but you drop them on your company blog with nothing to back it up, sounds like you’re the one who cares the most.
August 24th, 2006 at 2:51 pm
I am a financial analyst, so what I do is listen to yoohoo and yobnob types with titles like CEO, CFO, CEAWESOME, whateve talk bs and numbers at me and put it altogether through various complicated testing and then I make big money.
If I were to hear this:
“I could break it down for you, but there’s no reason to do so. The data’s accurate. ”
From a CFO, I would be about 95% sure he was completely full of sh*t and would stay away from his company.
We use several of the same stat packages you use, and we don’t see that kind of discrepancy. Blogads, Yahoo Publisher, Google Analytics and Sitemeter resolve to the same levels give or take a few %. Algorithms are for things like reach, unique visitors, etc. A pageview is pretty much a pageview, especially if your claim of culling the rss agents out of the stats is true.
August 24th, 2006 at 2:52 pm
“Dropped” by request of applicants. Yeah, we care. We care about people knowing we’re not a fly-by-night network with no income potential for great bloggers who’re applying.
Feel free to poke around, btw, and see how many PAN ads per page we’re serving. Because you couldn’t just leave it alone and believe me. Again.
I told you the stats didn’t matter. I told you we don’t care. The only time we talk about them is when we have to, because they don’t matter to us.
It’s not some big “pages served” counter on the top of the blog. It’s not something we report on monthly. When we passed 5, 10, 15, 20M pages per month we didn’t even *mention* them on our blog. When we passed more than a million uniques (and every milestone since) we didn’t even *mentioned* it on our blog.
Because we don’t care.
I do care about people who don’t know anything about our business, our servers, our infrastructure, our stats or stats in general calling us liars, though. Especially while trying to make out like they’re doing the world a service.
Next time just ask.
It’s one thing to call us to task when we make some big announcement. It’s another when we mention something in passing.
And when I say we do the traffic, don’t say we’re lying.
August 24th, 2006 at 2:54 pm
Johnny: If you, as an analyst, asked me, you’d get a different response than when Mike, as a competitor, says we’re flat out lying because of inaccurate data.
I have a responsibility to investors, analysts, principles, bloggers, partners, advertisers. I have no responsibility to a competitor.
August 24th, 2006 at 2:57 pm
Jeremy, you’re not a competitor. Like I said in the entry, we haven’t considered ourselves a blog network ever since we realized copying WIN wouldn’t be innovative, and that was a very long time ago.
August 24th, 2006 at 2:58 pm
Johnny: This is also why the world has comScore data, which is something we’re looking at. Because even in-house accuracy… Well, advertisers much prefer an actual third party doing outside auditing, just like financials.
August 24th, 2006 at 2:59 pm
“We’re not a competitor.”
Sorry. A company who’s traditionally gotten into it with us.
August 24th, 2006 at 3:00 pm
Haha, well shit man, I get into it with lots of companies, doesn’t make us a competitor
August 24th, 2006 at 4:35 pm
[...] I‘ve decided to take a quick break from my Making Code Poetry series after reading this little flamewar. In response to it, I have just one thing to say: Who really cares? [...]
August 24th, 2006 at 5:05 pm
I’ll post more later because I’m still in a client meeting as I do this - but jeremy is correct, BNL is out of date by nearly 60 days in terms of stats content. I will go run the script now to update the stats.. but the sites making up the networks are out of date by close to 60 days..
Matt
August 24th, 2006 at 5:17 pm
JDub,
I use statcounter now while i cant trust it 100% if you are going to step up and start saying we get 100gazillion pageviews you better be able to step up and show it. So far your numbers as an advertiser who has advertised on b5 across a few different sites I can say I’m a touch concerned with the accuracies and non accuracies. Just step up and correct the numbers I’m sure you get that many. I just think numbers are like getting into a fight over how many girls you have slept with. It’s pointless and very hard to prove. Cuz their isnt that grade a standard in the girl industry and in the stats industry.
August 24th, 2006 at 5:18 pm
MIke,
BTW Pat is an upright guy. RightMedia handles my PopCrunch advertising and I have to say I love it.
August 24th, 2006 at 6:38 pm
Traffic results should be read as if you are an accountant. It’s not important if they are right or wrong, but free from bias and consistent with other periods it is comparing to.
See wikipedia .. concept of MGAAP.
It’s like the concept of NET INCOME. If one year is on a cash basis, and the following is on an accrual basis - you can’t compare the two years because it’s like comparing apples to oranges. Net income is not a number .. it’s a state of mind (and subjective when it comes to income taxes - ask your accountant or CPA)
If Jeremy says last month they did 15M and this month 20M .. it doesn’t matter to me if they really did 6M last month and 8M this month - they still did 33% better.
Now, if you want to compare apples to apples .. why not compare eac other’s computer network system? or each others’ automobiles? or, if your dad can beat up his dad? or how much revenue your company did last month to his company? Then it’s a comparison.
But really, as fascinating as it would be to ‘really’ know .. it wouldn’t change my life in any way knowing or not knowing what the stats are .. but as always, I did get something interesting out of all these comments! Entertainment.
August 24th, 2006 at 8:14 pm
Mike, buddy, your newfound desire to “call out” people comes off as a tad obnoxious. Humblest opinion. But have fun with it.
August 25th, 2006 at 12:57 am
New found? Where you been man, I’ve been doin this for years.
August 25th, 2006 at 8:54 am
Indeed he has. The bulldog of 9rules.
August 25th, 2006 at 1:27 pm
Ah, just seemed like you were more snarky than usual haha.
August 26th, 2006 at 10:19 am
I’m wondering what kind of stats score you’d use for bareknuckle boxing.
BNL is way under on its assessments to the point of being a menace to the businesses it’s tracking. I’ve mentioned this to Matt and hope he might put an intuitive element in there to counteract a Technorati that might not have indexed you for six months, or an Alexa that’s weighted toward geeky websites.
But wait, isn’t that what Jeremy is saying?
August 26th, 2006 at 4:06 pm
Just want to butt in coz I saw a Media Kit with complete stats from one of B5’s blog: http://www.timedfinals.com/timed-finals-digital-media-advertising-information-kit/
I looked at the other blogs in the network and they didn’t have the same Media Kit page.
August 28th, 2006 at 10:42 am
[...] Going public, Mike Rundle has seriously questioned the numbers behind the rise and rise of b5media. [...]
August 28th, 2006 at 11:09 am
Mike, I hesitate to say this because I think you’re a hell of a talented guy.
But maybe the reason this blog doesn’t have more traffic is because you don’t focus on adding value to a larger audience. Sure, maybe the 9Rules choir enjoys it when you pick fights with b5, but don’t you want readers outside of that?
No one else care about this kind of stuff. And it’s tedious.
And yes I’m a boring old fart who should mind his own business. Just seems to me that with your skills you could be doing great things with the obvious large amount of time it took to prepare for, write, and excessively comment on this post.
Is this comment considered asshatting?
August 28th, 2006 at 4:31 pm
I don’t understand how asking a company to prove and explain their public statement is wrong. I read your blogs and you bloggers do it all the time. Many of you just did it when Business Week whipped up a number. In my opinion, B5Media (not Jeremy) made a statement (one their corporate blog) more than once about the millions of pageviews the company has. I don’t see anything wrong with questioning where the pageviews are coming from. Besides the owner’s blogs, none of them are run away hits. None are on the popular blogs lists. None of them have high traffic amounts. As a non-blogger what is particularly odd to me is how most bloggers (when they reach high traffic) openly say it. That is more than Rocketboom, TechCrunch, GigaOm, Scripting News, Robert Scoble…I could go on. As a reader that is a lot of pageviews to have very little presence on the internet.
I seriously doubt there would be this much backlash if a blogger asked the question - because it’s a valid question. The backlash is because 9rules mentioned it, which is a bunch of crap. If stats aren’t for open discussion then don’t talk about them although I can’t imagine why they wouldn’t be since the company depends on advertising to sustain.
August 28th, 2006 at 11:18 pm
I’m interested in learning more about stats packages. At Know More Media we’ve been using Sitemeter (which you can currently see on all our blogs - sidebar Sitemeter shows individual blog stats, bottom of page Sitemeter shows entire network stats). We’ve also used Google Analytics. I’m just curious to learn more about how to understand the differences. Maybe someone can help me try out AWStats or another stats tracker on BusinessBlogWire.com and see how the numbers compare to Sitemeter or Google Analytics.
Carry on :).
August 29th, 2006 at 12:28 am
Excellent post Mike. I realized that I should start jacking up my advertising prices based on my traffic. $50 for a 1 month text link is getting robbed considering my 5-7k pageviews per day.
August 29th, 2006 at 9:13 am
What I got out of this conversation is this: You guys care very little about readers, the public. By your own statements readers don’t matter, advertisers do. To most of you guys readers are simply numbers on a chart.
Maybe if you concentrated more on readers instead of advertisers you wouldn’t have to worry about fudging numbers.
This conversation showed me a lot about blog networks. It’s all starting to look like a big scam to get clicks and advertising money. I think as more people realize what’s going on and how little readers matter, there will be a backlash.
Thanks for opening my eyes.
August 31st, 2006 at 6:00 pm
[...] Professional Blogging Aug 31 at 6:03 pm by Thord Hedengren -Mike Rundle of 9rules has written a post on traffic, and b5media’s traffic in particular. It’s gotten a little heated in the comments, pies flying all over the place, but it’s still a good read for everyone that’s been thinking about stats and what they actually mean. You won’t find any truths, but if you didn’t have any doubts regarding stats before you’ll definitely get some. [...]
August 31st, 2006 at 8:03 pm
If I was saying I was the 3rd biggest blog network in the world and had 20million page views, I’d want to offer more than Jeremy has. This isnt about Mike its about supporting, OBJECTIVELY, with a “source” the facts being presented. Heck, this is first year university stuff. Maybe bloggers are the new Arts graduates training for flipping burgers, coz i dont see no blog network flips here.
August 31st, 2006 at 10:01 pm
[...] Professional Blogging Aug 31 at 6:03 pm by Thord Hedengren -Mike Rundle of 9rules has written a post on traffic, and b5media’s traffic in particular. It’s gotten a little heated in the comments, pies flying all over the place, but it’s still a good read for everyone that’s been thinking about stats and what they actually mean. You won’t find any truths, but if you didn’t have any doubts regarding stats before you’ll definitely get some. [...]
September 1st, 2006 at 3:27 am
Sorry this conversation has made you see all blog networks in the same light. I think some important points have been raised in this discussion, but not necessarily in the most helpful way. Not every network is all about the money, though, and I actually don’t believe that 9rules or b5media are purely about the money either - but unfortunately the advertising dollar often has an important say in what happens.
September 1st, 2006 at 6:20 am
Roy, you make a good point there. Blog networks are beginning to revert to type. The geeks who invented and run them are centered on stats rather than readers, SEO and content pages rather than providing great and varied reading material for normal people.
That’s one of the reasons Syntagma spurned the BN tag and now bills itself a Web Network Magazine. Read more.
September 1st, 2006 at 10:11 am
[...] I read with interest the debate going on about traffic numbers. [...]
February 17th, 2007 at 8:31 pm
[...] read more | digg story [...]
August 1st, 2007 at 11:11 am
[...] This quote is from an exchange involving Jeremy Wright, the President of b5media. B5Media is a large blog network. This quote is part of an exchange about claimed traffic levels to B5 Media blogs. The conversation is from 2006, but it certainly drives home the point: [...]
December 23rd, 2007 at 2:30 am
[...] Is Digg having trouble selling ads? [...]