Millions of Readers: Some of us have them, some of us don’t.
on July 30th, 2010This is such a long response, I thought I’d do me a favor and just make it a new post.
Follow up to a comment from the post below:
To clarify, the post says nothing about needing to get millions of readers a day to have millions of readers. That’s like saying the only sales a publisher gets are the sales it made in one day.
Quantcast tends to trend low on its numbers and shows that DC Comics.com gets about 200,000, and Marvel.com gets 1.2 million unique readers per month in the US alone.
So it is highly likely they DO have millions of readers. Neither DC nor Marvel’s numbers at Quantcast include global traffic.
I always read Quantcast with a jaundiced eye because its info about my site is way off, and it even has me linked in to WebComics World. I’m not. So I read multiple sites to get info and then average out.
7Zoom is the closest to my internal data (but today it’s way off, showing me at half my actual traffic. Go figure.)
I get about 5,000 hits a day, but that doesn’t mean I have 5,000 readers only.
I get somewhere around 200,000 hits per month.
Not every reader pops in every day.
So if I get 5,000 hits in a day, what I actually got was something like 1500-2000 readers that day. I may have a completely different set of readers the next day. My exact numbers for yesterday were 5863 hits and 1873 unique readers. The unique readers for the previous day were 1082. It is 9:35 AM on Friday, and we have had 712 readers.
My stat counter shows my unique hit count per month at a low of 38,000 and a high of 54,000. That’s misleading, because it is based solely on cookies and the way the system resets itself.
Stat counters use cookies to track info. If you are using a proxy or anonymous server, that skews data. Some systems reset info every hour, counting a unique reader who comes back as a new reader an hour later.
My core returning readership is actually about 10,000-12,000 people. They come back a few times a week. It’s far more accurate measure of true readership, not hits or unique readers. I would happily conflate my readership to include people who pop in once and leave, but that’s not real, is it? A hit is not a reader. I won’t do myself any favors by lying about how popular I am.
If you claim “millions of readers” on your print comic, and you get 1500 hits a day on the web, and your last book sold less than 3,000 copies through Diamond, you’re lying through your teeth about having “millions of readers”.
If you claim millions of readers on your webcomic, and you get 150,000 hits a day, and every stat monitor tells me you have 50,000 unique readers per month, I doubt you have millions of readers. What you have are tens of thousands of people who come back repeatedly every month. Which is fine. But 150,000 hits per day doesn’t add up to millions of unique readers, unless most of those hits is an individual reader, and that’s not the way it works.
100,000 devoted readers who come back once a day for 30 days adds up to 3,000,000 hits in one month. That’s nothing to sneeze at, and I’ll take it, thanks. But it’s still not millions of readers. And since only a small percentage of online readers ever buy, it accounts for the modest sales of some web successes. Of the 100,000 online readers, less than 10% will make a purchase.
The site I mentioned which claimed 3000 unique readers per day was boldly lying and knows perfectly well how to read stat info because the person running it claims to work on websites for a living. Their Alexa ranking is well under 1 million, and their (now defunct) Project Wonderful ads showed their unique traffic in the hundreds, not thousands. I went to the site to check it out as a potential place to advertise ADS, but I didn’t see any point in paying those prices for ad blocks on a site which gets significantly less traffic than mine.
A lot of people confuse hits with unique readers.
Torsten saw the 4400 number on my 7Zoom stats as my daily hit count.
“Yay! I’m one of the 4400!”
That is page views per day, not unique readers. That’s where the confusion about the size of readership comes from.
I work in the wonderful world of interactive advertising. Most agencies of any size have entire departments devoted to analyzing site statistics and most of the work we do where I work includes some site usage analysis work.
I would love to hear more from you on this subject.
Of the 100,000 online readers, less than 10% will make a purchase.
I’m one of those less than ten percent. If I like a webcomic I will buy the print version (my roommate got a bunch of those for me at SDCC, I sent her with a list). If there is no print version I will tip money their way, like I did last year sharing that unexpected windfall among a few favorites.
People look at me like I’m throwing money away. “Why pay for it when it’s right here, online, for free! DUH!!!1!”
To which I say, “because if I pay them, they can make MORE. Duh.“
Where I work, we have a Strategy & Analytics group that does nothing other than formulate and execute analytics plans. Ideally, they get involved at the planning stages of any site work deciding on both tracking goals (e.g. what site interactions we care about) and implementation strategy.
The analytics folks usually make use of various analytics tools that track users throughout their visit to a site. The analytics can get specific enough that the analysis team can tell on a page by page basis what parts of the page get interacted with more than others and how users typically enter and exit a page.
All of this may seem excessively detailed, but site editors take this information seriously and use it to shape how they develop their site further.
@Justin — Hey, do your analytics folks use Omniture/Site Catalyst?
I actually have found that advertising through project wonderful on sites with less page views isn’t always bad, especially if the taste in the site is close.
They are often relatively cheap, compared to other sites. (And to be honest, if someone likes something that is thematically close to yours, but less quality, it seems more likely they’ll be drawn to your webcomic. ;P)
What’s more, some of the sites with lots of hits (As well as some with bad hits, for that matter) have horrible ad placement.
I usually tend to follow the link to make sure I can actually see the ad when its placed, and I pay a lot of attention to how many people actually click the ad.
Sometimes, for whatever reasons, some sites just click better with your ads then others, and that can have little to do with page views.
@ Aurora: no it isn’t but if the site has a minimum bid on their ad box that is way out of proportion to their traffic, I will bid elsewhere. I’ve gotten great results from those little 1 cent per day ad boxes elsewhere.
You make really good points. The ad placement is crucial. Who care what traffic a site gets if no one can see the ad?
Do I count as one of the 10% if I bought all the comics BEFORE I ever found this site?
“Stat counters use cookies to track info. If you are using a proxy or anonymous server, that skews data. Some systems reset info every hour, counting a unique reader who comes back as a new reader an hour later.”
This is especially true with an IP such as AOL. Every time you log out and log back in it issues you a different IP address. Not that I used to exploit that flaw when voting for certain online comics in polls or anything (looks the other way, whistles)
Also, logging in from the same computer using a different wireless signal, such as free in the library or coffee shop, can register as unique as well. There are so many different ways to be counted so many different times.
Numbers don’t lie, when you actually read them. It’s when you interpret them that the trouble starts…
@VT – It depends on the client, but Omniture along with some custom tools made by the company are all in the toolchest.
It’s very surprising that using the same computer on a different network would make a subsequent visit to a counter or poll look like a unique visit. Using IP addresses as a measure of unique views just isn’t that reliable.
Using IP addresses isn’t a good way to track uniqueness because users will be coming through a proxy which will make multiple users all appear to come from the same source IP address. This isn’t just the case with people coming in from ISPs like AOL – most corporate networks are also set up this way.
I have a fixed IP at home, but I have to pay extra and have to have a business class connection as well.
Sorry, the above should read “most users will be coming through a proxy”.
Hooray for (not) proofreading.
@ Justin: Full agreement. I just had a long private conversation with someone over this. His counter measures unique IP’s and that is how he counts his readership base.
Using that logic, I have millions of readers, too.
@OnyxHunter! LOL! And welcome to the board.
@colleen: Actually, if he measures unique IPs, he probably has fewer readers, rather than more.
Sorry that I wasn’t clear.
Then again, I am not sure I am clear about how he explained his stat system. It really made no sense to me.
Basically, even though he realizes his readership is in the tens of thousands, he believes it is OK to claim nearly 20 million readers because he has had 20 million IP hits over the years.
That’s what I understand from his notes to me, but I respectfully disagree – assuming I understand his explanation correctly.
OK, here’s some math.
If your stat counter resets to count unique reader hits once a month, here’s how you can reach 20,000,000 without having 20 million readers.
20,000,000
Divided by 5 years.
4,000,000 hits per year.
Divided by 12 months.
333,333,333 hits per month.
Divided by 30 days:
11,111 hits per day.
That’s breaking it down simply without accounting for traffic increase curve, but getting 20,000,000 hits on a counter which resets unique readers only once a month can really inflate your numbers.
If it resets every hour or every day, you can end up with a ridiculously skewed view of how many people come to your site, if you are not inclined to critical thinking.
Since I have no idea how his system works, I can’t say for sure what’s going on. But his direct, measured traffic does not show millions of readers, but tens of thousands – and I’d be grateful for that.
I know what you mean by unique IPs not being unique readers.
I know I have accessed your site from
My home computer
My work computer
A hotel’s internet access while travelling
My brother’s computer
A friend’s computer last week in San Diego
My Droid – not sure how that one gets counted, especially when I am roaming.
So 1 reader and at least 6 unique IP’s.
Exactly. I must have accessed this site from dozens of locations and IP’s.
I am pretty used to the locations and IP’s of some of my friends and regulars.
My stat system will log each person on a single day, and show me where they have been and how many times they came back. For example at the top of their info, it will show their return visits that day.
Since I know where you live, and I know you are the only person from that city who has made a comment on the site, I am certain you are the person near the top of my stat counter right now.
I know you have been back more than 100 times.
However, because stat counters reset the unique IP info, it shows you have only been back to this site eight times today, and that today is your first visit.
When you come back tomorrow, it will seem as if you have never been here before. Your IP will be in the system as a new, unique visitor.
That’s another reason why the IP counter is bogus. It resets after a prescribed period. Sometimes as short as an hour.
The only IP’s my system tracks are those I program it to keep, such as an occasional weirdo who doesn’t need to be posting on this site.
Let’s see… sale of ADS = 40,000. Let’s multiply that by the number of issues = 1,560,000. All of those readers go and buy the four graphic novels, so that’s 6,240,000. And, let’s assume a pass-along readership of three: 18,720,000. There you go: near as damnit twenty million readers. Easy.
It won’t be long ’til JK Rowling comes knocking at your door pleading poverty.
If only all my issues had sold 40,000! LOL!
Assuming best case scenario on audience base is pretty ridiculous, as you demonstrate. That is exactly how some people conflate their readership.
I just tried to edit this and lost half the post. Oh, well.
All the sites visits talk can be really confusing. When people ask me about how many people visit my blog it all gets confusing and I have nowhere in space and time near the amount that visit your site.
With all that said half my pdf sales are word of mouth and half through random searches on LuLu.
My main blogger site, my main site, doesn’t get that many at all compared to my livejournal; which I’ve had for a long time and another one for even longer before this current one. My facebook gets a lot more than my livejournal but hardly anyone goes from there to buy things as far as I know. I get more emails and DM on Twitter so unless you are me you would think no one cares. Not even counting where sales come from my brain can go up or down depending on how I see and count things. I have to be careful or I will think I’m doing well with traffic on a site, when in fact I just used the right dirty word one too many times
Oh, and when I comment on an entry I come back many times here to see if anyone commented on my comment thus but that’s because there is no system set up to send me an email when a thread is update…or is there?
I honestly don’t know if readers can get emails to see if they got a comment, because I only see from the mod end.
To illustrate the IP problem, if you use AOL, every single time you click on a page, AOL may assign you a new IP.
Someone who decides to sit down and read ADS all the way through will show up on my stat counter as dozens and dozens of AOL users from the exact same town looking at only one page of the comic, and each with a different IP.
That doesn’t happen every time, but a lot.
This is why you can’t just look at your biggest number and assume that’s true. Be critical. Pushing your raw data forward as an absolute is done either from ignorance or a stat conflation. It’s the same reason publishers add up all the sales on a book and claim their million fans.
If you sold 100 comics at 10,000 each, you sold a million books. You can tout your million sale, but what you really had was a core audience of 10,000 readers, not 1 million readers.
If your book sold 100,000 copies in 1987, but sold 2,000 copies in 2007, it’s misleading to say you have 102,000 readers.
What you had was 100,000 sales in 1987, and an audience that shrank to 2,000 in 2007.
Of course no one wants to say that in public.
‘Since I know where you live, and I know you are the only person from that city who has made a comment on the site, I am certain you are the person near the top of my stat counter right now.
I know you have been back more than 100 times.’
I did not know I was that transparent. I was hoping to be more of an international man of mystery.
I generally check your site daily and if I post a comment, I check back periodically to see what replies are made.
I don’t think the site is set up to send emails when there are replies to comments.
LOL!
If anyone doesn’t want to be seen, they can just use an anonymous proxy, and some people do.
I don’t do anything with the info except monitor the system for traffic trends and malicious activity. If I don’t know you personally, there is no way I could match the info to an individual reader.
Awhile back, someone wrote me to tell me that he was afraid that if I knew who he really was, I could have him blacklisted. He often wrote very unflattering things about me on the board, and I caught him using sock puppets (a ban-able offense – don’t do it.)
That’s just daft. I can’t get a stranger’s true name from their IP.