Peace talks held over autorefresh rates for websites
The explosive confrontation between website owners and media agencies over the use of autorefresh to inflate page impressions appears to be closer to a resolution after the key players met to discuss the issue.
Although the question of how often it is reasonable for a site to refresh a page has been a live one for years, the debate has reached boiling point in recent weeks.
Those who defend the use of auto refresh point to the fact that it is reasonable to refresh some pages often, such as news home pages and latest sport scores.
But media buyers argue that they can end up paying several times to reach the same reader on the same page, without there being transparency about it.
Anyone who pays per impression is a fool, but the publishers won’t change this ruse until the advertisers stop paying for it.
Can’t news websites implement an AJAX-style solution to only refresh the certain sections of the sites (such as homepage news feeds etc) so that the content stays up to date? Google does this with Maps, their live search results and so on.
Hell even using the old iframes would allow them to do this.
Content on a view should be able to be updated without the advert being refreshed if need be. it just takes some tweaks with the frames and the coding – it should be possible.
As for hitting the same user twice – I do not understand what is wrong with that? If I place an advert in a magazine I am sure that the same person will read that mag a fair few times and see my advert again and again, just like they will drive past the same billboard and see my tv advert time and time again in between the cricket, soap opera etc
Do news homepages need auto refresh though? Why not have a big red ‘hot’ button on all news sites that becomes uniform, which reads: “click to refresh” – job done…
I’m with Simon and James: but perhaps this is too costly an option to implement portal-site wise AJAX.
On the flip side, until media agencies stop paying for it, they publishers have no critical reason to implement it.
Agree with both comments.
He who pays per-impression deserves everything he gets – literally!
And of course there are technical solutions to the this. But why would you want to fix the problem when you still have the means and opportunity to fleece your advertisers?
Does this remind anyone else of Copenhagen? Perhaps there should be an autorefresh trading scheme.
#3 Fresh: Part of the problem with autofresh (besides driving up the number of ad impressions served, when the user hasn’t taken any action to change the page) is that autorefresh stops sessions timing out. Therefore an autorefresh-using publisher can claim that users spend longer on their site than they actually do, even if that person just left the homepage open, refreshing for hours. Longer session times = higher CPM since you’re reaching a “more dedicated audience”.
Solution is to charge a flat daily rate for home pages that auto refresh. Publishers are loath to implement AJAX and i-frame solutions as they screw up their internal metrics and have a negative impact on SEO.
Ajax – the technique web developers use to retrieve information without reloading the entire page – can’t be ignored in the autorefresh debate.
Websites will increasingly have Ajax-driven dynamic interfaces that will mean readers will find what they want but in doing so they will generate fewer pageviews in the traditional sense.
News Ltd sites get roughly 30 to 40 per cent of their hits from auto-refresh on a 4-minute cycle. I’m very surprised they’re not on the list !
“He who pays per-impression deserves everything he gets – literally!”
So erm, how should advertisers be paying then?
And please don’t say CPC or CPA. Please.
Re auto-refresh and the excuse around ‘keeping news most current’ … not sure I agree. Many article pages auto-refresh … why is that? So we can get the most recent edit from AAP??
Can’t help but think auto-refresh is less about keeping content fresh and more about increasing pv’s/inventory to sell per unique. Plus it games ‘engagement’ metrics as well which really clouds things for advertisers.
I’m glad something appears to be being done but at the same time sceptical about follow through from many after these initial talks. Would love to see a load of green ticks next to Australian sites on MI within 2-3 months. Will we?
Why stop at auto refresh……what about the publishers (umm, 3 of them) that default you to their homepages when you log out of your email service, surely this practice is also to inflate homepage traffic numbers.
Take those out and see how big their homepage numbers are??
The thing that’s always annoyed me about these large autorefresh sites is that these sites and/or publishers are frequently (perhaps by sheer coincidence) the same ones that dont offer or easily accomodate frequency capping.
Capping by U/User views, whilst I acknowledge that its not without its shortfalls atleast offers some haven against the merciless CPM churn of autorefresh.
with all the questions around the data/numbers most publishers put out there – and the reluctance to be completely transparent – it will be a great day when there’s a uniform set of guidelines for all sites to adhere to.
whilst we’re at it, we may as well address the trend of publishers (such as The Age and SMH) auto starting video in articles (many of which are just still photos with a voiceover) … videos that add nothing to the article.
the cynic in me says they’re trying to boost their video impressions as the yield on these placements is far better than straight CPM but maybe there’s some overriding noble reason I’m missing. Whatever the rationale … as a user it’s annoying and I’m not the only one who thinks so.
Ben: “So erm, how should advertisers be paying then?”
The problem with auto-refreshed page impression numbers is they’re _meaningless_. We have no idea how many of them have actual eyeballs looking at them and how many are someone stepping out for lunch with the SMH home page open. The alternative is to advertise only with properties that don’t auto refresh. There’s definitely a place for display advertising that is paid by impression on an eyeball rather than PPC — branding exercises really.
Until the auto refreshes go away, you’ve taken the world’s most measurable advertising medium and made it less measurable than television’s decidedly dodgy measurements. Advertisers shouldn’t put up with it, but while they still do I wouldn’t suggest anyone use CPM display on these sites.
simon – the problem with digital’s self promotion of its ‘measurability’ is more often than not this premise (false or otherwise) is abused. familiar with measurement for all media i’d say digital display is right now THE least measurable and easily the most flimsy when analysed.
Re not paying CPM. Most other approaches can be equally manipulated. Paying per unique. Paying for a click. Paying for an action. Set a scenario for revenue and unfortunately many will spend the majority of their time trying to game it.
sorry cut myself off above … meant to say that for the above reasons the ABA’s work should be a priority for all involved, so look forward to seeing the next steps from todays meeting.
CPM is fine if the publisher is ethical. Many publishers do not have auto refresh and they have a very good audience. The big publishers need to be monitored and a uniform code needs to be introduced. An ‘ABC’ style clear, transparent audit of all websites. Could the IAB get involved with pushing something there?
The folks advocating AJAX content refresh are misunderstanding why publishers put it on there. It’s really got NOTHING to do with refreshing content, and EVERYTHING to do with inflating page impressions, session length stats and so on. Publishers don’t really care about showcasing new content as it is published — as that would only generate a small/incremental number of new impressions, whereas autorefresh delivers vast numbers of impressions due to people’s habit of opening many tabs in their browser and leaving pages open for hours at a time. I’m surprised at the person who said News Ltd gets 30-40% of impressions from 4 minute refresh. That frequency of refresh would actually deliver 200 – 250% more impressions than they really have.
would be interesting to look at average ad response (be it clicks, interactions) over the past 5 years and see the trend
with auto refresh and clutter becoming more prevalent, what is it doing to ad response/impact?
The problem is not just with publishers – it’s ill-advised advertisers signing up for bum deals and end up paying exhorbitant rates for questionable outocmes.
Advertisers – typically at the advice of their media agency – use impressions or reach as some kind of measure of success. This is bad. If it’s not driving traffic to your site, why bother?
If it’s driving traffic cost-effectively it doesn’t matter if publishers are auto-refreshing (well, yes, it’s wrong, but not all publishers do). If the effective cost per visit gets too expensive drop the CPM deal with that publisher and reallocate spend elsewhere eg. search, seo or cpa.
For many years, Anthill’s website anthillonline.com avoided using auto-refresh for all the reasons identified above. Auto-refresh is like printing 100,000 newspapers, citing this as circulation, then throwing 70,000 in landfill.
There is another issue that has not yet been raised.
Nielsen Online Ratings charges its clients according to PIs pro rata – it earns more revenue from organisations with higher PIs. As such, it has a vested interested in not baring auto-refresh bandits from its auditing system, like the ABA has done.
The only way that Nielsen can support this policy is if it overhauls its pricing structure, to perhaps charge clients according to UBs instead?
Leon – i think it’s a bit limited to say that driving traffic to an advertisers site is the only objective worth measuring results against.
Page impressions? Is this article a reprint from 1997?
On weatherzone the auto-refresh is only enabled on local weather pages, radar pages and pages that carry weather warnings. This is prefer most users prefer these pages to auto-refresh as it keep them current – same argument as sports scores. Our auto-refresh is 10 minutes because this is the update frequency of radar and current weather data.
However in the next version of these pages we will give the user the option to switch auto-refresh on or off. This will restrict auto-refresh to those who truly want it and should address advertisers’ concerns.
Auto-refresh is a topical issue internally here as well (weatherzone). During big weather events, the auto-refresh drives bandwidth use to unsustainable levels. On these days we have on occasions turned auto-refresh off in order to maintain a decent quality of service. Yes, our PIs per visit are lower when we turn it off but not dramatically so.
My personal feeling is this feature should only be enabled when the content warrants it (news headlines, sports scores, weather warnings etc.). But having run our site in both modes for some time I don’t see it as a significant issue from a metrics or advertiser value point of view. We will always focus on the user benefit as the primary factor in making this decision.
Apparently… and this is unsubstantiated water cooler talk… Australian click through rates are low in comparison to international rates. The person who shared this factoid with me attributed it to Australia’s love affair with auto-refresh. I don’t know whether it’s true and whether auto-refresh is more prolific in Australia than elsewhere. So, if someone can offer some data to back this claim I think we’d all be interested.
Q. Is auto-refresh usage more prolific in Australia than elsewhere?
Q. What benchmarks do other countries use to sell ads? (PIs also?)
Q. What auditing standards do other auditing bureaus demand of their members?
The big issue here is credibility.
Media buyers are responsible for trying to spend client money in the most effective way they can.
So it is in the online advertising industry’s interest to provide meaningful PI and UB (frequency and reach) numbers – especially if the media owners/networks want to get repeat business.
Counting for the big three in Australia has always been first about the business and second about the specific display opportunity. Not that dissimilar to other display media if you are familiar with the arcane rules and disputes of print, radio, tv and outdoor.
As always in this discussion we have one set of publishers (smaller ones) complaining and digital agencies trying to protect the good name of the digital opportunity.
Any agency planner should be able to deliver their clients value with any site including any auto-refresh site. It’s part of what they do.
Standard ways of talking about counting are all good and have been around for years. If your planner can’t get the information they need from the big three then there are lots of opportunities elsewhere. Easy as.
Sites will always be innovative in white and black ways that will mess with some mistaken belief that all websites should be able to report their opportunity in comparable ways. Of course they can and of course they can’t or won’t.
re the 40% News Ltd figure. That data is based on figures from one of the main NDM sites. After a recent revamp the auto-refresh was accidently not switched on for about 10 days. Page impressions down approximately 40%.
It’s quite depressing to see how much of this discussion is about publishers, advertisers and their all-important “metrics” and SEO and how little — almost none! — is about whether any of this auto-refresh povides value to the audience. Where is the audience’s need being discussed?
Sorry. I just realised what I said. What am I thinking? As you were.
Stilgherrian: publishers don’t care about what the audience wants, as long as the stats are good! They only care about the stats they can sell to advertisers (as long as they are not pissing off the audience so much that it might do damage to future stats.) In publishing, readers=stats.
If you want an update on the pushing and shoving that is always bubbling below the surface in online measurementland try this and the attached comment stream.
http://www.techcrunch.com/2010.....unch-face/
At the end of the day it’s only about display opportunities and only about big vs small. There is no one size fits all golden standard that will solve this, never has been and unlikely to be, it’s knowing what is being bought and each on it’s merits.
The debate in the US re Comscore now is very interesting … seems like a lot of those involved in slating Comscore want to move 100% to server side measurement and want to trash the accuracy of panels.
Which is odd – considering server side logs are often equally unreliable – counting bots, not accounting for cookie deletion, not accounting for users home/work duplication, or dual browsers, or intl visitors etc.
Have to wonder whether the Comscore discussion (ie charging non subs $10k to include them in the more robust measurement pool) is more motivated by sites wanting to build credibility around server side or a legitimate whinge about a $10k cost.
Same issue is active here as far as I am concerned. Some sites quote MI (site side) and others quote Netview (panel) … which one should agencies believe (I use Netview personally for all numbers but not sure everyone else does)
Bill, regarding your comment:
“Stilgherrian: publishers don’t care about what the audience wants, as long as the stats are good!”
I hope that you don’t work in publishing.
We avoided auto-refresh because we found that it’s annoying to the audience. Who wants to be half way through reading a story and then… zing… you’re sent to the top of the page. Or half way through writing a comment and then… zing… your thoughtful prose is gone. These are things that readers hate.
As mentioned above, Anthill scored ABA’s highest session time in November without auto-refresh. We scored that accolade by caring about what the reader wants.
But here’s the kicker…
Over the summer break we added auto-refresh as one of our many experiments (expect the full report soon). We set our site up to refresh every 10 minutes (while our competitors refresh at 3 mins). Our session times jumped fourfold!
We were already leading the pack and suddenly we had session times four times longer than our nearest competitor. And that’s just session times.
We’re about to crunch the numbers to see how PIs stack up but you can expect that it will be a similar story.
I said it above, I’ll say it again (but revised to reflect the News Ltd statistic offered by Smithee). Auto-refresh is like printing 100,000 newspapers, citing this as circulation, then throwing 40,000 in landfill.
Jerrys – With so much money at stake, a clear standard will emerge (whether it’s ‘golden’ or not, is another story).
Quoting James Tuckerman:
“I hope that you don’t work in publishing.”
I’m just telling it how it is. Real estate agents don’t care if tenants like the properties, as long as they keep renting them. The customer is the landlord, and the product is the leasing fees.
In publishing, despite what they teach you at journalism school, the customer is the advertiser, and the product is advertising space.
Note, I work in editorial, but the reality is that editorial doesn’t make a cent by itself — it’s advertising that pays the bills. And advertising is sold, partly (or largely) through positive stats. Hence my comment that publishers don’t -really- care about what readers want, as long as the stats continue to be good.
Perhaps I should have been clearer and said: commercial publishers only care about what the readers want as far as it contributes to positive stats that they can take to advertisers. That’s why we don’t see the kind of stuff that goes to air on SBS and ABC as much on commercial TV channels.