Is a view the same as a hit ?
A hit for those of you who dont know is merely the presentation on a webpage of a file by a webserver. You put 10 graphics on a webpage, you get 10 hits when that page is displayed.
Internet oldie goldies will remember the entertainment from the mid 90s of sites publishing and talking joyfully of how many hits their websites got.To the uniformed, of which the business and general media was a card carrying member, references to hits were meant to represent how much traffic a website got. Stories were written daily referencing the “number of hits a site got” as a way to represent their traffic. Of course it was absolutely not the case. Fortunately these days, its rare you hear people use hits as a traffic metric.
Today though, with all the excitement about rich media, whether its video or audio, you hear and read all about views and listens.
Which leads me to the question: Just what is a view or listen ?
For podcasts, is a listen a download ? I hope not, i have several podcasts in Itunes i subscribe to, and that download every episode, but I havent listened to in a long while. A couple others I have listened to the first several seconds or minutes of, but havent really dented. How would what i just did fit into the podcasts metrics ?
In radio there is Time Spent Listening. There are quarter hours , where a radio station gets credit for someone listening for 15 minutes if they listen for 5 minutes. Are there standards for podcasts ?
Then there are videso on the net. Just what exactly is a view ? Our HDNet Trailer for an upcoming Dennis Rodman show has been viewed 2339 times. I know i have gone to the site and it has started running at least 20 times that I didnt watch it all the way through. Are those counted ? I cant find anywhere on the Youtube.com site that defines how views are counted. Did i miss it ?
And this isnt a reference on Youtube. Its a question for an industry. Streaming video servers and Ad Insertion servers can count exactly how many bytes were streamed to a user which can be recalculated into number of minutes , seconds or hours for each attempt at viewing a video, whether ad or content. And from experience, i can tell you that a huge chunk of attempts to view are aborted for any number of reasons. Yet I have yet to see any references to views with an average length of viewing or a percentage of video watched competely through.
Just what is a view ? I have a feeling that its pretty darn close to a hit these days.
BTW – your commenting system is going crazy – it\’s sending me multiple \”confirmation emails\”, yet I click just once and my comments are posted several times. It\’s not me trying to rain on your parade – really 🙂 Rob
Comment by ehliyetsnavsonular -
I bet you your hot items will be sold out soon. With its protecting system, your business interest will be guard against those business scams, you wonder like me where to start your business on Call2biz.com, the Call2biz.com is your source of detail information need to do it. We would help you and marketing your products without any charges of fees! You will save up to 100% and earn more easy money on Call2biz.com. with a only computer and phone at your home, you can, while being successful, enjoy this easy-money-making experience online. A homebased business will begin at http://www.Call2biz.com.
Comment by call2biz-blog -
As a videoblogger/video-podcaster, I have to say that most of the time a hit is an actual view. If the video has been downloaded, the chances are very high that it has or will be viewed. Here’s my reasoning…
-iTunes policy: When speaking with the producer of Tiki Bar TV last night he told me that a contact of his at Apple actually mentioned that iTunes only downloads up to 5 podcasts automatically before cutting the client off. iTunes will not start downloading any more new content until those podcasts are viewed (or listened to).
-Podcasts vs. Video-Podcasts: While Podcasts are usually something you don’t actively pay attention to, videos usually demand your full attention… so unless it’s a really boring video (or a video that might as well just be audio because there’s no real visual stimulation), it’s probably been fully watched.
-Size: If each episode of a video-podcast is around 40MB, and someone subscribes to the show on a weekly or daily basis… they will eventually suffer from a clogged up hard drive. This will either lead them to unsubscribe or just delete old shows. If the subscriptions of a podcast go down, perhaps this is why (either way, it’s relatively trackable).
-YouTube: When you visit their site, YouTube counts a page view as a video view. They do this because the video starts automatically. However, if you embed the player into your blog/site/etc. YouTube only counts it as a view if you click to view the video.
I think a larger issue on this topic would be total hits vs. unique hits. On our Blip.tv statistics we can see that nearly everyone who watches our show, watches it twice. It’s probably due to our geek culture, I know that I personally love to re-watch my favorite shows.
Thanks for checking to make sure the comments were working, Mark! Take care…
Comment by Casey McKinnon -
You can actually count the ads as they are delivered. No one uses the “I sent the ad and so someone must have seen it” method. IAB guidelines are very clear on the method of counting ads.
Comment by Ele -
No hit.
Comment by Jastrzebia Góra -
First… Google Analytics blows for site metrics.. install the server version or urchin if you like the way it works…
best program that is not a server log (and works better than server logs…
http://mcc.hitslink.com
oh… the is B.S.
You must have a view to have a hit.
You have to have a visual to make a hit.
You have to have vision to have a hit. IP
and image that is 1px by 1px that loads on another page can constitute a hit. The 3 buttons for adding an RSS feed on the upper right hand corner of this page (yahoo!, newsgator and feedster), all push hits somewhere… the best way to track anything is uniques, page views should go down on any website with better penetration on SEO..
As for your question… simply bust out a refresh or a browser open via java at the end of the video, use flash and run the video within the flash and at the end of the sequence, open a new window and track that page load/ or the unique visit to that page if you want to get single viewings… multiple viewings won’t show up because the script won’t run twice due to most pop-up blockers or a unique event counter in your flash program, unless of course they change IP addresses between viewings… if u want to chat about it.. give me a buzz mark.. just Google “Paisley Does Dallas”
Comment by paisley amoeba -
First… Google Analytics blows for site metrics.. install the server version or urchin if you like the way it works…
best program that is not a server log (and works better than server logs…
http://mcc.hitslink.com
oh… the is B.S.
You must have a view to have a hit.
You have to have a visual to make a hit.
You have to have vision to have a hit. IP
and image that is 1px by 1px that loads on another page can constitute a hit. The 3 buttons for adding an RSS feed on the upper right hand corner of this page (yahoo!, newsgator and feedster), all push hits somewhere… the best way to track anything is uniques, page views should go down on any website with better penetration on SEO..
As for your question… simply bust out a refresh or a browser open via java at the end of the video, use flash and run the video within the flash and at the end of the sequence, open a new window and track that page load/ or the unique visit to that page if you want to get single viewings… multiple viewings won’t show up because the script won’t run twice due to most pop-up blockers or a unique event counter in your flash program, unless of course they change IP addresses between viewings… if u want to chat about it.. give me a buzz mark.. just Google “Paisley Does Dallas”
Comment by paisley amoeba -
A View is A Hit === if you view it as a hit, it’s a hit, otherwise it’s NOT a hit… everyone has a differnet view (or Hit) depends on their needs..
bobby
Comment by Bobby Orbach -
Hey mark, if we can “sort of” tell how long someone visted our site via a Cookie. Why can’t a video plant a cookie and monitor time? … Then compare it to how long the video is, then you could have another catagory instead of “views” have one say “Avg length Watched” and it would be a percentage like 71% and then someone could immediatley see how long others watched a video … You could also take the end number like “Number of Views” Times your “Percent completly watched” … Then you have the number of “Whole times watched” … I hope my thinking is right, Thanks ,
-Richard BOwles
Comment by Richard Bowles -
Hello, see this page – Links Fight . Com – it’s an awsome idea!
Comment by Matthew -
Sorry Justin, as much as I dislike Mark Cubans, he’s right. Congrats on building websites since 1994 and never actually knowing what a “Hit” defines!
Comment by Alex Grosholz -
“A hit for those of you who dont know is merely the presentation on a webpage of a file by a webserver. You put 10 graphics on a webpage, you get 10 hits when that page is displayed.”
Am I the only one who saw the inaccuracy of this statement? Granted, I did not read all fifty-some comments, but the ten or so that I did read all skipped this point and delved into the podcast issue.
But I’ve been building websites (albeit none are successfull) since 1994 and found this claim to be grossly incorrect. Simply having ten graphics on a page does not earn you ten “hits” after a single visitor loads the page. It earns you only one. And if your server software is anywhere in the vicinity of current, multiple visits from a single IP are NOT logged as multiple “hits.”
The file which earns you a “hit” when loaded by a visiting netizen is only the index html document, not the individual contents within that page.
Comment by Justin Kolenc -
“A hit for those of you who dont know is merely the presentation on a webpage of a file by a webserver. You put 10 graphics on a webpage, you get 10 hits when that page is displayed.”
Am I the only one who saw the inaccuracy of this statement? Granted, I did not read all fifty-some comments, but the ten or so that I did read all skipped this point and delved into the podcast issue.
But I’ve been building websites (albeit none are successfull) since 1994 and found this claim to be grossly incorrect. Simply having ten graphics on a page does not earn you ten “hits” after a single visitor loads the page. It earns you only one. And if your server software is anywhere in the vicinity of current, multiple visits from a single IP are NOT logged as multiple “hits.”
The file which earns you a “hit” when loaded by a visiting netizen is only the index html document, not the individual contents within that page.
Comment by Justin Kolenc -
What if an office or school network has 5,000 people behind a single IP using NAT? our office use one IP about 30+ staff.
Comment by YAMADA -
I think Podcasts, RSS, Ajax and all the modernity that web2.0 has brought with it will be here for quite a while. We will see some client side applications like browsers getting a little more intelligent and aiding in the collection of data.
Web 2.0 may come with it’s own standards and a system for traffic analysis must be put in place eventually. The situation is quite similar to the transition from snmpv1 to snmpv2 – Trial and error and eventually the industry converges on the most acceptable standard.
Comment by Ken -
Do we really want that kind of detailed info and how do you keep it secure?
SEO and internet marketers still use the “hit sales pitch” on internet newbies daily. I get calls from forums from many scam companies that are selling garbage and many of them brag about how many hits they get on their sites.
Comment by port orange real estate -
A view is the same as a hit. You can figure that out by simply hitting the refresh button on the brower while at the viewing page. :-/
I know nothing about PODcasts.
Comment by Todd Johnston -
Do pod cast really matter? In your opinion is there really a beneficial market for them? Is anyone actually making any money with them?
All The Best
JP
http://www.sportsfansattic.com
Comment by JP -
Mark,
Your endless quandering always amazes me. Why would anyone think of these things? You do though and I feel as if your inquisitive mind is making me smarter.
am I kissing ass? Maybe. But at least I am learning about things I would never in a million years think about.
Thanks!
Comment by Greg -
Having signed a significant number of podcast advertising deals we have used some pretty solid performance stats when talking to perspective advertisers.
I would be happy to discuss how we lay out our advertising plans
Comment by Todd Cochrane -
it is bigest post what I can see
Comment by GlooM -
It’s a woot off Mark….a woot off!!!!!!!
http://www.woot.com
Comment by Matt -
For streaming comScore MediaMetrics reports all of the relevant information you are asking about. Time spent listening, unique viewers, total bytes transferred, average number of clips per session, sessions per user, and many additional categories.
As you are well aware the big difference between terrestrial radio/TV and the internet is that you don’t have to estimate viewer ship to determine how many ads are served. You can actually count the ads as they are delivered. No one uses the “I sent the ad and so someone must have seen it” method. IAB guidelines are very clear on the method of counting ads.
http://www.iab.net/standards/broadband_video.asp
A beacon is sent from the client viewing the streaming ad after the video is visible on the computer. Sure, some people may stop viewing the ad but it most sites won’t let you see the content until after the ad is played in it entirety.
People should not be satisfied when a site reports “views”, I agree this is a completely bogus measurement. Let’s talk unique visitors and total time spent viewing content.
Comment by Justin Madison -
Anyone who spends more than 5 minutes a week at youtube and is over the age of 16, needs to get a life.
Seriously.
Comment by Beef Jezos -
I believe the correct and more accurate statistic is the “unique visitor” cat.
Comment by Jacob Janik -
No doubt, web metrics are over-inflated no matter what measurement you go by and always changing.
Externally viewed RSS feeds also show as a unique visit, but how many people actually read anything from the feed readers or even see it on their my.yahoo page, or wherever, is questionable. And, if any graphics are involved, it can easily increase your number of hits dramatically.
Plus, all the measurements become skewed if your ISP uses dynamic IPs or if users share one IP as mentioned before.
Interpretation is an art as much as it is a science.
Comment by Brian Meyer -
No doubt, web metrics are over-inflated no matter what measurement you go by and always changing.
Externally viewed RSS feeds also show as a unique visit, but how many people actually read anything from the feed readers or even see it on their my.yahoo page, or wherever, is questionable. And, if any graphics are involved, it can easily increase your number of hits dramatically.
Plus, all the measurements become skewed if your ISP uses dynamic IPs or if users share one IP as mentioned before.
Interpretation is an art as much as it is a science.
Comment by Brian Meyer -
Mark,
This is pretty lame comentary for appears to be just a marketing spot for a Dennis Rodman show. The real question is how many readers here actually clicked on that link. The subject matter would have been fine, except it smells of stream of conscienceness thinking and adds no real value to the topic. You pose a question and give no real insight or your own opinion on what the right metric should be…
The IAB has already addressed this issue according to this article. http://www.iab.net/news/pr_2006_05_15.asp
In summuray, I enjoy your site and generally like agreeing or disagreeing with your thought processes, but this post just screems of trying to get all the yes men posters to click on a link to buy more popcorn.
Sean
Comment by Sean O'Keefe -
Well, one of the problems is that there aren’t tools for accurate reporting on actual duration of play, even for “live” streamed media where that info should be in the logs. We evaluated several of the major Web Analytics vendors’ products that allege to give streaming media stats and the results were laughable at best.
Comment by Ernest Mueller -
I think if you get someone who hits “play” that is a hit/view.
The company that drives the traffic to your server or charges you for adspace only gets paid to get your ad viewed.
Your message and the content, whether the message be for a good, product or service, is what is responsible for you holding the attention of an audience.
I say if you get someone to click, the ad did it’s job.
Advertising should generate more revenue then it costs, and if you can’t make that work, you’re doing it wrong.
Comment by Mark -
“Streaming video servers and Ad Insertion servers can count exactly how many bytes were streamed to a user which can be recalculated into number of minutes , seconds or hours for each attempt at viewing a video, whether ad or content.”
In a cached, streaming environment, bytes streamed don’t mean bytes viewed. The player needs to report the exit point. Even then, that doesn’t mean it was viewed. I regularly open 10 YouTube pages at once to get them to load up while I do something else. They each may play in the background while I’m away. Then I go back through them and watch from the beggining, but most don’t get completed.
Comment by Brad Murray -
I dont view Itunes as a good analogy -a hit is a hit there as its whether you download or not, your paying full freight for it. Easy to measure.
Comment by Brian -
http://adage.com/digital/article?article_id=111011
It seems plenty of advertisers share your opinion. Online video is simple to measure, unless it’s autoplay in which case you have to ask whether that’s a useful ad or an annoyance.
YouTube uses a very simple code to measure views; whenever play is pressed at the beginning of a video. A bad way to get that data since it doesn’t account for skipping or closing, but that data isn’t for anything useful. Calacanis made a good point on his blog that YouTube isn’t run like a business (see downtime) and their metrics reflect this.
As far as podcasting goes, I’ve never been sold on podcasts as advertising mediums beyond sponsoring specific shows but hey, it’s web 2.0! who’d overpay for a company with real statistics…
Comment by Adam Cains -
I’ve always said HITS stands for How Idiots Track Success.
I know for streaming audio it measureable how much people actually listen to the stream. We see reports on the average listening time which tells you how “deep” people get into the stream.
I’m not aware of a corresponding measurement for a podcast. I thought you just got subscription numbers for metrics.
Comment by Tim Taylor -
Is A View The Same As A Hit?
It depends on what dimension your analyzing.
They could be the same simultaneously, although one populates the other, as the result of the hit is the processing of the view.
You must have a view to have a hit.
You have to have a visual to make a hit.
You have to have vision to have a hit. IP
The metrics of a view calibrates the same as a hit, although there is interactivity with the hit that could calculate additional time factors/elements, as the view is a one time visit, indicating a visitor, with the traffic metrics while the hit can be several aspects in one for traffic analysis.
Either way an analysis of metrics should determine the multiple site interactivity thereby being a hit, as the view initiates connectivity, a vision of talent to a hit, a visual of site to make a hit, a view of a website page to a hit on the page.
Comment by S. -
I really don’t understand podcasts. They just talking, bla bla bla. And if I want to relearn information?! Where I will get it? Also visual memory is much more better than aural memory. So you better understand it from docs, then from mp3s
Comment by Tom Martin -
It doesn’t *really* matter as long as everybody is comparing apples to apples.
As to your podcast example — a download counts as a “listen.” Even if only (say) 50% of those downloads are actually listened to all the way through, it’s still an apples-to-apples comparison. Maybe there’s a range there. Maybe for some podcasts, 70% actually listen and for others only 30% listen. But I would think that (on average) the listen to not-listen ratio is pretty similar across all podcasts, and therefore it’s just a scaling factor across all podcasts.
So call it 50%. So if I’m comparing two podcasts’ popularity (say one gets 20k downloads and the other gets 10k downloads) then the actual “listens” are only 10k and 5k — but #1 still has twice the audience of #2. It’s still apples to apples.
Comment by Brock -
To add fuel to the fire, how do you determine unique visitors? Is each person tied to a single IP? What if an office or school network has 5,000 people behind a single IP using NAT? Tracking web metrics is by far the hardest part of running a successful site.
Comment by Chris -
What is a “view”? Depends who’s asking.
For advertisers, I would argue it’s the number of people who saw or heard their ads. TV advertisers are just getting to this, with Nielsen’s new “commercial minute” ratings. These ratings are imperfect but better than what was previously available.
For content producers, I would argue there are two relevant metrics: the number of downloads, and the number of people who saw or heard the end of their program.
Comment by ken -
Joe. Free publications use audited circulations. In addition, paid circ uses a Pass-Along rate to find out true readership.
Comment by clark -
Oh, and I completely forgot – yes, AJAX does change the face of the metric-game. Ads no longer need to be persistent – they can rotate, similar in delivery to the sideline-ads at basketball games.
Comment by anton -
Personally, I think you’re asking the wrong question. Everyone is so caught up in the numbers game of web-traffic metrics that they forget the purpose(s) of advertising – to generate revenue and build a brand for your product.
I rarely have time to listen to podcasts, but I’d be interested to know: are most ads ‘spliced in’ – meaning pre-recorded, produced, etc., or is it simply product-placement by whomever is doing the podcast? Obviously, product-placement has a much larger impact on the target audience, and I think that this is the way that advertising in vlogs is going to go. Yes, total visitors and other metrics are important, but how the ads are delivered is equally, if not more important.
Long story short – using short-terms metrics to study long-term habits of consumers is counter-productive. I know I’m going to get modded as flamebait for not offering any solutions of my own 😉
Comment by anton -
I agree with you Mark that a page view is getting awfully close to a hit, which is why I use Google Analytics on my site to measure specific goals and page conversion – you have to determine what the tasks your users are trying to accomplish, and then figure out how to get your website to help them to accomplish those tasks, and then benchmark where and when you can measure if those tasks are getting done. If you want to get downloads, how many “page views” does it take to get there might be a measure.
For instance, one goal on my site is to determine potential leads for legal cases, so I want to track page views on specific topics. For that goal, page view numbers are relevant, because I need to know which pages are getting seen and searched for so I can put that content together for my users in a way that will get them to call our company – another goal. But for tracking how many calls and requests for information I get, it’s more important to know what pages they looked at, and how many clicks it took them to get to the form, then how many pages, necessarily.
Comment by Michelle Tackabery -
Mark,
Great post! I’ve been working in the web analytics industry for many years now and and have written three books on the subject. It’s exciting when well-read people ask the really hard questions like you just did. I blogged about something very similar just a few days ago — the notion that “Web 2.0” will need to be meaningfully measured (original post is at http://www.webanalyticsdemystified.com/weblog/2006/08/we-want-web-20-measurement-standards.html)
While you describe one side of the problem (when is a listen a page view and when is a listen only a hit) I think Jeff (post 23) nails the other side of the problem: the emergence of AJAX, RIAs and externally deployed applications has the potential to quickly muddle all of the hard work that folks have done to measure the impact of their online presence, further confounding the online advertising issue.
The good news is that bright people using great technology are already out there trying to answer your question. Given the relative maturity in the analytics marketplace, I believe it safe to assume that somewhat banal questions about measurement will quickly give way to business questions and good Web 2.0 key performance indicators like “application event views per session” and “complete listens per visitor” driving online marketers to keep what works, dump what doesn’t and continue to optimize the overall visitor experience.
Eric T. Peterson
Author, Web Site Measurement Hacks (O’Reilly)
http://www.webanalyticsdemystified.com/weblog
Comment by Eric Peterson -
I think we should creating IE-Explorer and/or FireFox plug-ins that can measure hits and/or refreshes on the client site, and relay that data to the server. With all the upcoming technologies the idea of measuring from the server side will no longer be accurate.
Comment by Mitchell -
I wonder how this works out on MySpace also with the music players. It’s obnoxious hearing peoples attempts at music start playing every single time I go to there page. I usually make it a point to get to the music player and press stop or pause before it can begin as to save my ear drums the trouble of hearing their “music”. I bet it counts as a play though. Damn it! >;[
Comment by Luis Alejandro Perez -
Most site metrics track unique visitors, visits, and hits. I can see how many average visits a unique visitor makes to my site and how many hits per unique visit.
Comment by GolfersWired -
On a simliar note.
How do you count hits/impressions/etc for sites that utilize Ajax?
What about these Ajax sites if they make money off ads? CPM in particular. How much money are they losing because the number of ad calls have dropped b/c of Ajax (assuming they didn’t have it originally)? Could that be made up by addtional users?
Random thoughts along the same lines.
Comment by Jeff -
how many people buy a magazine and don’t read it? That counts as a sale. Same with newspapers. Neilsen ratings don’t just give a rating to shows that are watched all the way through. A person can turn on the TV, fall asleep and wake up 6 hours later. All those shows he slept through are considered views.
The only way that a site can truly be ranked is by the ability to sell merch. An entertainment site with links to amazon should be ranked by the ability to get people to click and buy the DVD or book that’s being raved about. Influence is all about how much money a person’s site can get flowing.
Right now my Youtube posting of my “Slacker With No Shame” segment from the Today show has 300 views. Are those unique views? Or is it my mom watching it a dozen times? Ultimately, I don’t care. I just like the idea that my friends around the world can see Campbell Brown talking about me without me having to make copies and mail them out.
Comment by Joe Corey -
they’re probably not calculating that information because it’s not in their best interest to do so. kind of like way back in the day, if newspapers didn’t need audited circulation, would they go thru the time and expense to see how many papers are REALLY being read?
Comment by clark -
Just use Google Analytics 🙂
Comment by Marco - Share Trading -
kr8tr hits the nail right on the head. Metrics data for online media offers such tantalizing (if false) promises of exact traffic and data, that we forget how inexact previous methods of measuring broadcast metrics are.
If you go inside the sausauge factory, so to speak, with someone like Media Audit, you see them telling online organizations they’re viewership in a particular region is greater thent he total amount of data they’re serving worldwide. I’ve seen this happen. Now, if the relationship continued, I have no doubt that the data would have gotten better. But a traditional organization – like, this thing you call “radio” – with no way of actually knowing if numbers are bullshit, just accept what they hear.
So, we’ ve got two sets of standard here. One, where you accept whatever fairy tell about non-traceable materials that you’re told that sounds good. And a second, where you expect to know to the second that people listen and follow your content. I think the second is an admirable and interesting goal, and the issues of people gaming nubmers and systems are not to be dismissed. But, lets not pretend the existing systems were fine, and these new ones are flawed. Both are non-ideal systems in a non-ideal (*cough* *privacy* *cough*) world. Cheers!
Comment by James Cooney -
I agree that the measurement for podcasting and videocasting is not exact. As we have seen through the evolution of internet marketing, technologies have allowed for stronger and more accurate measurement which increases value for marketers and raises the revenue for publishers (at least those who perform well).
The issue of podcasting and videocasting is challenged from multiple areas: offline aggregators/players (iTunes, Lemon, FireAnt, portable players, etc), indexes (Odeo, Yahoo Podcasts, Podshow, etc), and definitions regarding partial plays.
Aggregators/Players: The current method for determination is downloads. Currently, there is not a feedback loop established for any media that communicates plays to a central or originating site. As Apple is generally a closed company, we cannot expect them to open up data to the content publisher. There are attempts with plugins to iTunes that report back usage, however, outside of Acrobat and Flash, it is hard to say that plugins are a usable solution. A simple solution, if agreed upon, would be a reporting tag stored within the metadata (id3 tag, etc) for reporting a play. This could be built into any player and enhance all reporting, though it would have to be implemented into the aggregator/player by the software company.
Indexes: mp3 is generally considered the lingua franc of podcasting, and flash players are ubiquitous across most of the podcasting indexes. There again is a lack of reporting regarding playback of files, only downloading. From the above example, a simple tracking url encoded into the metadata (id3 tag, etc) could provide that feedback. In addition, partial plays could be reported to this source. One important feature for all indexes is to not autoplay – require the user to press play. This would define that the user has chosen to play.
It is also possible to throttle the file download on the server to match the encoding to limit the amount of caching for a flash player. This again is not exact, however, it will give a better view point, from the server, of which files received listens or the percentage of partials.
We have built our flash players to anonymously report back as it plays. This provides us with the ability to track actual listens. However, this does not provide us detail on plays from Odeo, Yahoo, Podshow, etc.
Partials: The issue of partial listens has not been standardized. With a webpage, you are dealing with fairly small downloads. However, with a media file of 12-80 MB, what constitutes a full impression? If you listen to 18 minutes of a 20 minute podcast/videocast, is that a complete listen? If the advertisement is pre-roll or interstitial, does listening to 50% of the episode account for a full listen?
Mark, I would agree that there are not enough standards within the medium. However, there are a bunch of us working very hard on solutions.
Comment by Brian Walsh -
*NOTE* – I said I had a URL – a domain name. There is no web site there worth looking at – I was just trying to make an example. The poddigger.com domain is spoken for – but there has to be others out there that are even more suitable.
Rob
Comment by kr8tr -
Mark, as always, an astute thought process pertaining to something us business entrepreneurs are looking at. Angles to get our business viewed. If someone sees a link or your name out there, you have made a connection and a familiarity with that brand identity. The more they see it, the chances are, eventually… they go to it.
Thanks again, and please consider buying the Kansas City Royals. 🙂
Comment by CultivateGreatness.com - Travis Wright -
It’s late, and I am sleep deprived – but Ken Carpenters post (two or three up from here)got me thinking about something I think might be useful here.
What we are really concerned about is “interestingness”. IF we can measure that, we can determine what is or isn’t worth advertising on – and even what we should pay to advertise. Flickr does it very well now with pictures.
We need a “digg” for podcasts. I have a URL – poddigger.com – a “social site” for podcasting. The site rates podcasts, by the people, for the people (and for the benefit of the advertisers as well).
If enough people think something is interesting, does it even matter if 10-20% of them never listen to a podcast they think is interesting? I think not – the web is a huge place, and aggregated data that says “This is interesting” won’t come easy. I imagine over time that the shear number of people “voting” on the “interestingness” level of a podcast, or video blog (or blog for that matter) will prevail – and advertisers will have found a targeted audience.
Comment by kr8tr -
Keep in mind that IP-based counters aren’t perfect either. There are a host of problems with accurately tracking the session of a user, especially if they don’t want to be tracked. Proxy servers make many people come from the same IP, people share computers (public terminals even), others reject browser cookies, etc, etc. Then there’s the whole world of click fraud.
Tracking could definitely improve in the video / Podcast world – but it’s worth noting that it will still suffer from the same problems as tracking any other kind of content.
Comment by Brian Yennie -
A downloaded podcast is exactly like a magazine subscription — it gets delivered to a potential reader/listener, but no one can prove how much the recipient listens or reads.
It’s an antiquated advertising medium. Unless there are toll-free phone numbers or e-mail addresses or URLs presented in the podcast or magazine ad, you don’t know who is actually reached by the sponsor.
Advertisers need to pay based on sales that can be tracked to the medium, not on the size of delivery audience.
Comment by Ken Carpenter -
If your counter doesn’t suck, it will count only unique hits/page views. Meaning, it will ignore an IP address once it’s seen it.
Comment by Chris Davis -
Julie,
Think of it in more narrow terms – there is no “mass market” for podcast listeners. There are pockets of listeners that each have their own devotion to x, y, or z. Some listeners are transient, and they walk away without ever “joining the cult”. Other listeners are fixated beyond belief on a particular podcast.
If you accept that as being true, then as a advertiser you know some important things: “What my listener is interested in” and “most of these listeners are interested in enough in this to deal with an ad”. Remember – there are fanatical listeners to some podcasts – and most of these are being made on reletively cheap equipment by relative amateurs. Imagine what a little more cash and experience could do to improve the experience (NOT suggesting that Advertisers take over directing podcasts – that would kill the fun for everyone!)
That means very targeted advertising. Targeted advertising means a more substantial return on the advertising dollars spent. Podcast advertising is already dirt cheap – and if the market can find a measurable way to ensure you pinpoint your advertising then advertising suddenly costs you less, and rewards you more. You have basically just increaded your gross margin.
Comment by kr8tr -
I work on web metrics every day. According to my definition of a “view” on streaming media is this:
A view is counted upon the initial play event of pre-recorded stream. If preceded by an END event, subsequent play events are also considered views.
I’m not sure if every site counts it like this, but that’s how we count streaming media for my clients.
Comment by Dan -
Thanks for bringing this vital subject to the attention of all, Mark. We’re producing a series right now on “Monetizing Internet Video” featuring industry thought-leaders, innovators who are pioneering new models, along with big media’s dominant players in today’s market –ie, Nielsen. Next week we have a short video piece up on our website from NAB featuring interview on this very topic of METRICs. It’s a great piece, but not entirely fulfilling because so much has yet to be settled on industry-wide. I honestly I find it astounding how little progress has been made on this front. Until this issues of standardization is fixed industry-wide, the VOD potential of the internet will remain unfulfilled.
Comment by Megan Cunningham -
Mark:
I have the same questions also. I have the idea that views are logged instead of hits because the number is bigger and therefore gives the impression that traffic is greater. I do not think the difference is substantial between the two. I want to think that the difference between inches and centimeters is more substantial than the difference between visits vs. hits vs. listens.
Here is a link to an article attempting to explain it:
http://ezinearticles.com/?Page-View-Logic-As-Applied-To-Article-Writing-and-Marketing&id=81907
Comment by Antonio Howell -
Maybe I just have the attention span of a gnat. The podcasts were accumulating in my library and I just gave up.
Comment by Julie -
BTW – your commenting system is going crazy – it’s sending me multiple “confirmation emails”, yet I click just once and my comments are posted several times. It’s not me trying to rain on your parade – really 🙂
Rob
Comment by kr8tr -
Julie,
I don’t agree – pod casts are NOT over-rated. I do agree that they are under-measured. That’s a different problem. I know pod casts haven’t exactly hit every iPod – but neither have banner ads ;).
Pod casts have a lot of value because they capture attention. Anything that captures attention *should* have value to advertisers.
Rob
Comment by kr8tr -
The Pod Cast can be mesured in increments. 100 increments being a full and completed.
If it stops at number 50 increment you have to determine what you want that to say to you. Count it as a full or consider it a drop out and consider it a hit.
If you get a 51% complete, you have to make your rules as to read it as a drop or a full. You have to set your measuring increments to say what you want to know about the cast.
Comment by BG COP -
Right now, placing ads on podcasting and internet videos are *the* big things online advertising. (I sell advertising for a fairly large website and we get requests for podcasts and video all the time.) And you’re right, you have no idea if podcasts are ever listened to or if videos are ever seen the whole way through. So many sites are throwing the newest web 2.0 gadgets on their sites because they are in demand, but they may not be fully developed or completely right for their site. What is my point? Measurable media, such as banners and text links, are much more reliable and podcasts are overrated.
Comment by Julie -
I agree completely on this, since I’ve struggled with the same issues in some of my consulting work. Specifically to pod casts, I think there are only two ways to actually measure “listenership” – one is to create a new standard for players (or a new player, or both) that “talks back” to the original site with information like when it was downloaded, if it was listened to, and when, and how much played. Not an easy task.
The second way, and MUCH less accurate (I know – I was a “Nielson Family” for years, and I know I was grossly inaccurate) is to poll listeners.
Neither solution is a solution – and this is the dilemma of our times – there are too many ways to access content, and no real way to compare the relative value of a commercial I watch on the web vs. one I watch on TV.
Some genius out there may claim to have a solution to this, but I would almost blindly (without even seeing their proposal) dismiss it. There isn’t a solution. This is just “Something New” – and how we deal with it, market it, profit off of it or even measure it is completely new territory. I don’t think there is “a solution”. I think there are partial solutions. I think there are unique ways to re-engineer things like pod casts and videos so they (at certain timestamps of playing them) ping-back to some centralized server, basically saying, “Hey this dude got this far”.
Comment by kr8tr -
I agree completely on this, since I’ve struggled with the same issues in some of my consulting work. Specifically to pod casts, I think there are only two ways to actually measure “listenership” – one is to create a new standard for players (or a new player, or both) that “talks back” to the original site with information like when it was downloaded, if it was listened to, and when, and how much played. Not an easy task.
The second way, and MUCH less accurate (I know – I was a “Nielson Family” for years, and I know I was grossly inaccurate) is to poll listeners.
Neither solution is a solution – and this is the dilemma of our times – there are too many ways to access content, and no real way to compare the relative value of a commercial I watch on the web vs. one I watch on TV.
Some genius out there may claim to have a solution to this, but I would almost blindly (without even seeing their proposal) dismiss it. There isn’t a solution. This is just “Something New” – and how we deal with it, market it, profit off of it or even measure it is completely new territory. I don’t think there is “a solution”. I think there are partial solutions. I think there are unique ways to re-engineer things like pod casts and videos so they (at certain timestamps of playing them) ping-back to some centralized server, basically saying, “Hey this dude got this far”.
Comment by kr8tr -
And on the flip side, if I really like the music in a podcast or am really bored at work and listen to it 5 times, how does that get calculated.
With things that get downloaded through lets say iTunes and then moved to a recorder, I’m not sure how they can measure anything beyond downloads. Plays would be skewed because only a certain percentage would actually play the media on a device that could report back any metrics.
Comment by Mike -