{{ searchResult.published_at | date:'d MMMM yyyy' }}

Loading ...
Loading ...

Enter a search term such as “mobile analytics” or browse our content using the filters above.

No_results

That’s not only a poor Scrabble score but we also couldn’t find any results matching “”.
Check your spelling or try broadening your search.

Logo_distressed

Sorry about this, there is a problem with our search at the moment.
Please try again later.

Few topics in online advertising generate more confusion and debate than view-through attribution, and the debate isn’t likely to end anytime soon.

Marketers today are typically including it as part of the overall measuement of their ad campaigns, but still have questions about how to measure its validity, what percentage impact on consumers it is really having and how it is overlapping with the other marketing channels.

In Chango’s newest white paper, “View Through Attribution Exposed: What last touch isn’t telling you”, we’ve tried to put the confusion to rest. Properly used, we believe view-through is a valid metric that can help brands understand the true value of their display campaigns. Let’s start with the basics.

View-Through Basics

Most of us see display ads every day, but few of us click on them. A study by comScore and Starcom MediaVest found that 84% percent of Internet users never click on an ad. But just because you didn’t click on an ad, doesn’t mean that ad didn’t enter your consciousness. Maybe later that day, or the next day, you found yourself visiting the site that was being promoted in the ad. That’s a view-through visit, or post-impression visit. If you performed an action while visiting the site, that would be a view-through conversion.  If you purchased a product while visiting the site, the revenue would be view-through revenue.

How can an advertiser know that a visitor to the site has already seen a display ad? Let’s take an example. Say you see an Amazon ad while reading The New York Times online. When the Amazon ad loads, it will drop a cookie on your computer’s browser. When you later navigate to Amazon on your own, Amazon can recognize the unique cookie generated by its ad.

The Case Against View-Through

The above description of the view-through metric probably doesn’t sound very controversial. So why all the fuss about view-throughs? The primary complaint is that view-throughs, unlike clicks, don’t provide a direct line between an ad impression and  a visit to a site.

As this line of thinking goes, the Internet was supposed to turn advertising from guesswork into a science. Every marketing dollar would be clearly accounted for because every user action could be tracked via clicks. View-throughs, while providing an indirect line from exposure to an ad to arrival at a site, still leave room for uncertainty. If a user sees an ad, and visits the site in question the next day, there’s a good chance the view-through drove the site visit. But, in most cases, it's impossible to be certain that any given impression led to a visit. 

View-through skeptics can also point to the fact that it wasn't long ago that many media owners were attempting to rig the system by buying millions of cheap impressions across the Web in order to drop as many cookies as possible.  Fortunately, this practice, known as “cookie stuffing,” has grown increasingly rare as the industry has matured and grown more transparent. 

And then there’s still the question of the view-through window, that is, how much time passes between the moment the ad is generated and the moment the user visits the site.

If a user sees an Amazon banner and then goes to Amazon.com three weeks later, should that count as a view-through? Unfortunately, when view-throughs were first adopted, the standard was to set the window to 30 days, a length of time that is far too long for most campaigns.

Based on our experience, the 30-day window continues to be used by most marketers today. We suspect that if the more skeptical brands and marketers recognized that the window can be adjusted all the way down to 24 hours, they might change their minds about the usefulness of the view-through metric.

The Arguments For

Despite the reasonable concerns expressed by the view-through skeptics, most marketers today use view-through measurement as part of their digital marketing programs. This probably doesn’t come as a huge surprise. After all, the majority of ad dollars are still spent offline, and offline advertising essentially relies on the view-through model.

When advertisers run offline campaigns, whether for TV, radio, print, mail, or outdoors, they rarely have definitive proof that customers are responding to their campaigns. Brands understand that these campaigns — whether for branding or driving a direct action — work because they correlate with increased conversions. It’s not a perfect science, but it works — and has worked for a very long time. 

While there are some ways to verify view-through in the offline world, such as with coupon redemption, unique codes and selective DMA targeting, most offline campaigns lack even these assurances. Online view-through, by contrast, is more direct and accountable than most offline campaigns, because, at the very least, brands can be confident that their ads have been seen. The key is in understanding the right way to measure the efficacy of the view-through campaign.

View-through proponents also point out that just because clicks can be counted doesn’t mean they should be counted. The vast majority of Internet users never click on an ad, and the number who do is dwindling by the year.  

Meanwhile, those who do click rarely convert. Clicking is generally random and doesn’t correlate to actual on-site conversions. In fact, view-throughs account for over 90 percent of site visitors and will be responsible for over 90% percent of page views when they get there. In other words, view-though turns out to be the more reliable metric in the end.

Another good reason to trust in online view-through measurements is that their effectiveness has been studied and is now well documented. By comparing samples of users who were exposed to a displayed ad to those who never saw it, marketers can see the incremental value from the exposure and determine the revenue such exposure accounts for.

Search data also offers plenty of evidence of the value of view-throughs. Countless studies have shown that individuals who are exposed to a brand’s display ads are much more likely to search for that brand and click on that brand’s search ad.

Though the effect can’t be directly measured, there’s good reason to think that display ads can make a big difference in in-store sales as well. According to one comScore study, 82% of online ad campaigns led to an average lift of 22% in CPG brand sales in retail stores. 

The Verdict

Counterintuitive though it might seem, the action of a click is irrelevant in display media. Views-throughs might not provide the perfectly efficient advertising model we all hoped for as the industry developed, but the model has worked well offline since the birth of advertising. And the online one works even better, both because we know the user has been exposed to the ad and because there are a wide-range of studies that demonstrate the correlation between online view-throughs and visits, search traffic, and revenue.

As with most things in life, the key to recognizing the value of view-throughs lies in choosing the right tools and using them properly. Fortunately, we now have plenty of tools and the days of arbitrarily assigning 30-day view-through windows will hopefully soon be gone. A

midst all the numbers, there’s another point worth remembering when you think about view-throughs. Sometimes it’s okay to use your gut in addition to your analytical mind. After all, if the view-through model had no value, there wouldn't be an advertising industry today.

Dax Hamman

Published 6 August, 2012 by Dax Hamman

Dax Hamman is the Chief Revenue Officer at Chango. 

2 more posts from this author

Comments (19)

Comment
No-profile-pic
Save or Cancel
Avatar-blank-50x50

Ed Lamb

I'd be interested to know the source of this stat and how it was arrived at?

"view-throughs account for over 90 percent of site visitors and will be responsible for over 90% percent of page views when they get there"

about 4 years ago

Peter McCormack

Peter McCormack, Founder at McCormack Morrison

It is funny that this report is by Chango with their vested interest. Any digital planner worth their money knows full well that display marketing performs very poorly. It is a brand channel at best for large spend brand campaigns.

The only people keeping it alive are media buyers who might as well work for the Syrian authority with their ability to tell the truth.

about 4 years ago

Avatar-blank-50x50

George Symeonidis

I'm a bit skeptical about how we know the user has actually seen the ad in the online environment (the author says "online works even better, because we know the user has been exposed to the ad").

In offline, when the TV or radio is playing an ad, that's all it's playing. We can measure how many people were viewing/listening during that ad break, so we know how many people the ad reached. Granted, some of them might not have been paying full attention, but they were exposed.

In online, when a display ad is shown right next to the content the user is reading/interacting with, how can anyone be sure the user even saw it? I don't have any stats at the ready about ad blindness, but as I said above, I'd be really skeptical before taking for granted that a user was actually influenced by an ad just because it loaded on the page he was visiting. Not to mention ads that load below the fold, but I'd assume modern ad serving systems take this into account.

On the other hand, if stats show it works...

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

@George - that's a very valid point, and not one we shy away from. There are some tools emerging that help us (as an industry) measure not only above and below the fold, but what percentage of the ad was on screen and whether it was moused over etc.

Like TV though, we can't tell if they were 'out of the room'.

The only reasonable way to test any of this is with some sort of benchmark test and measure incremental uplift. Until tracking gets better of course.

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

@Ed - fair. The stat is simply observed, either from prospecting programs we run today, or from when I was on the agency side. Happy to discuss more offline if you are interested in data.

about 4 years ago

Peter McCormack

Peter McCormack, Founder at McCormack Morrison

The reporting of display marketing is a joke. I am yet to see a single company report a successful campaign based on clicks, hence view through, the coincidence metric.

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

@Peter Why do I feel this isn't going to be a short back and forth :) Controversy seems your MO for getting the debate started, so sure, I'll jump in.

Ignoring the Syrian references and unaccepting viewpoint that people who believe in VT could be correct, there are some interesting points you make.

Firstly, there are reports that show campaign performance can work on the click, and some site retargeting and prospecting buys can be purchased in such a way. The problem is in a lot of cases the click is just invalid, as shown in some of the reports the white paper links to. I particularly like the study where Quantcast demonstrate the randomness of clickers.

So the argument goes that with the click being questionable, we need to find a way to measure what impact display has. Advertising works, and this might not be a topic for you if you don't believe it does. Offline has proven it time and time again. So we can say that digital advertising works, the debate being more about the best way to measure it.

There are of course many ways. I personally prefer either an awareness / brand study such as a Vizu report, or to link the display impressions directly to revenue generated. With the latter we also ask clients to take the order IDs we generate and see if any other digital channel has touched that person, and from there produce a model of how much likely contribution we made whilst the program was running.

The problem in the industry is that we try and track too much. We look for direct and transparent links that might not exist, it doesn't mean to say correlations don't either.

about 4 years ago

Avatar-blank-50x50

Domenico Tassone

Bravo Chango. The white paper is a really good attempt to address one of the biggest challenges in measuring display media: passive response.

As an analytics practitioner and veteran of the digital ad industy, the fact is that viewthrough is a very valuable metric albeit often misunderstood. The reality is that it requires both technical capability and an understanding of basic experimental design...too often that is challenging to do.

Peter: while I can appreciate the skepticism, your assertions end-up being very weak. Simply because you have not had the ability to measure display properly does not mean that the vehicle cannot work; in the right hands it can be very powerful. You can educate yourself more about viewthrough measurement here:

http://www.tipofthespearblog.com

about 4 years ago

Peter McCormack

Peter McCormack, Founder at McCormack Morrison

View through was invented when clicks demonstrated an ROI way below search. Speak to Expedia, they dropped their display £10m display activity and saw minimal impact upon sales.

And as for retargetting, it really is just a way of claiming credit for a sale which would likely have happened anyway.

The easiest way to smell a rat is the way someone has to write a report every couple of months justifying display. The only people who like it are those who make money off selling it to clients who don know better.

about 4 years ago

Avatar-blank-50x50

Deri Jones, CEO at SciVisum.co.uk

Dax

Peter may be blunt, and hey that's often good on an online forum to trigger more people to take part!

But his challenge to the honesty question does challenge some of the things you've written:

> There are some tools emerging that help us (as an industry) measure not only above and below the fold, but what percentage of the ad was on screen and whether it was moused over etc.

So you're saying those tools are not in use today right, for today's campaigns?

George asked:
> how we know the user has actually seen the ad in the online environment

Dax, can you point to any scientific studies that say how much/little of a web page is not observed when a user reads it?

After all, the human brain is almost designed as an 'information filtering' tool to help us ignore the 99.99% of visual + aurual info presented the whole time!

I'm sure others here have seen the science lectures on TV where a room full of people totally don't notice a gorilla walking through the centre of the stage, because they'd been asked before to focus on something the people at the front are doing !

And a lot of UI time is spent trying to ensure that users spot the features we want them to spot on the page - knowing that too easily users will totally miss 'obvious' links.

So - some real scientific studies on how much of a web page gets missed, would be interesting in this context.

(it's one of my obsessions Dax - evidence based decisions)

about 4 years ago

Avatar-blank-50x50

Chris Brinkworth

@econsultancy. Having the ability to vote comments up / down based on a readers fair judgement of a comments value/input, would be a useful addition to your site moving forward.

about 4 years ago

Peter McCormack

Peter McCormack, Founder at McCormack Morrison

Good idea Chris

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

@Deri Great feedback and questions.

Firstly, I am not sure which commentators have read the paper or not, so I want to make one thing really clear - I do not believe in 100% view-through, far from it actually. But nor do I believe in 0% given that we can show advertising can influence people. All we can do therefore is answer the question of how much of it is valid, and for that I proposed a few methods, but I accept there might be better ones (that exist today or not), and that each method has a degree of error.

With that said, there are some tools that look at above the fold versus below the fold, but they aren't commonly used. At Chango for instance we use some that help us define the situation, and some media exchanges pass this information, but it is far from common usage yet. That does make measurement harder.

I don't have access to the data you are requesting to hand, but I did work in usability too and so am sure I can find something appropriate. Watch this space, and thanks for joining the debate.

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

@Peter - re your comment:

>And as for retargetting, it really is just a way of claiming credit for a sale which would likely have happened anyway.

I agree! Site retargeting is seen as an acquisition vehicle when it fact it should be measured as a conversion tool, perhaps more as a sunken cost.

Note that we focus on search retargeting, which is actually a prospecting vehicle that doesn't talk at all to existing site visitors.

about 4 years ago

Matt Owen

Matt Owen, Head of Social at Econsultancy

Hi @Chris

Yes agreed. We have been looking at various possible additions to the blog functionality recently and votable comments has come up a couple of times. No timescale in place yet I'm afraid but it is on the agenda.

about 4 years ago

Avatar-blank-50x50

Matthew Lindenberg

The thing about viewthrough conversion is that its incrementality can, in fact, be measured - I've done it. Take the audience pool that would be targeted for the ad, and as you see them on the internet, randomly segment them into A and B cells. Serve cell A the intended ad, and serve cell B a public service announcement. Tag them, and track the conversion rates among both groups who saw the ad, didn't click, but still converted - and measure the difference in conversion rates.

When I've tested this, I saw a lift among those served the ads, but it was only about 3% - not really a powerful statement in support of viewthrough.

about 4 years ago

Dax Hamman

Dax Hamman, Chief Revenue Officer, and digital geek at Chango

Matthew - interesting, that does happen. What % of spend were the control ads versus the main campaign? And did you include or exclude site retargeting from the mix?

about 4 years ago

Avatar-blank-50x50

Matthew Lindenberg

The test had 80% served the ads and 20% held out. It included retargeting.

about 4 years ago

Avatar-blank-50x50

Alex Baker

Interesting article and comments here guys.

I’m new to the world of online advertising and have recently joined a retargeting company who works on post click sales. One thing is confusing me – if as an advertiser you can work (and pay) on a post click attribution model, why would you pay on post view sales?

I appreciate that PV enables you to gauge the branding effect of a campaign and also enables you to benchmark a campaign against other forms of advertising that have to work on a PV basis (i.e. offline). However, from what I understand, there are some fundamental flaws with PV as have been pointed out in this article which haven’t really been addressed:

• View-throughs, unlike clicks, don’t provide a direct link between an ad impression and a visit to a site.
• Post view is still, to a certain extent, guesswork not science.
• With post view, it's impossible to be certain that any given impression led to a visit.
• “Cookie stuffing,” has grown increasingly rare – has it? What evidence is there of this?

There are also some interesting comments on working on a post click basis:
• The vast majority of Internet users never click on an ad, and the number who do is dwindling by the year.
• Those who do click rarely convert.

As I mentioned above, I work for a retargeting company that works on post click sales model. We do not charge our advertisers for post view sales (although we do record them for reference). This company also happens to be the largest and fastest growing retargeting company in the world so I can’t see how the above is true. Maybe I’m just biased…

As an aside, Google Ad Words does not work on a post view basis.

Thoughts welcome!

about 4 years ago

Comment
No-profile-pic
Save or Cancel
Daily_pulse_signup_wide

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Daily Pulse newsletter. Each weekday, you ll receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.