“Banners Don’t Work”
Brian Morrisey recently wrote about a conversation with Eyeblaster’s Dean Donaldson about why ad execs don’t like banners.
the notion that banner ads don’t work is mostly [due to] poor creative.
Morrissey is one of the smartest journalists in the business but I was disappointed that he didn’t dispute the underlying premise that boring banner ads don’t work.
The “banners don’t work” meme has been a quiet, industry conventional wisdom since the dawn of online advertising. Jason Calacanis complained back in 2001:
The banner ad is a complete, unmitigated failure: When we look back on this industry, it will be considered the biggest mistake we ever made.
Ironically, Calacanis made a small fortune by creating and selling, Weblogs, Inc. — an online publisher that earns most of its revenue by selling banner ad space to marketers.
When Supply Exceeds Demand
In an earlier post, Banners In Crisis, Morrissey invokes AOL’s decline in ad revenues as evidence that the format is going through a crisis.
Display advertising, if you want to use the fancier name, is in crisis. AOL just reported an 18 percent drop in ad revenue – and that’s with growth in search. Its display business is so bad that AOL this week bragged about a deal to run FreeCreditReport.com banners all over its ad network. It’s not alone: Yahoo and Microsoft both have struggling banner businesses.
It’s arguable that publishing is in crisis but that probably has more to do with the underlying economics of the industry rather than a flaw in an ad format. There’s just too much supply.
The biggest challenge to publishers is selling inventory at premium CPMs (e.g. $10+) when there’s a glut of comparable inventory selling at CPMs closer to $0.35.
Overvaluing The CTR
Another cause of the belief that banners don’t work is our collective over-emphasis of the value of CTR.
Yes, the average CTR is low and continuing to decline. But CTR should never have become an important metric for marketers to focus on.
- It’s not normal for human beings to click on ads. Human beings, whenever possible, tend to avoid ads for the distractions they are. Low CTRs should be expected.
- CTR has always been low. Throughout its history, the average CTR for normal ads has hovered well below 2%. Today the average CTR is close to 0.10%. Marketers focused on CTR tend to miss 99.90% of the value they create.
- There has never been a study that has shown a positive correlation between CTR and brand awareness or favorability
- CTR is rarely an appropriate metric for campaign effectiveness except for pure DR programs where revenue per thousand impressions and revenue per click are measured in tandem with CTR.
What About Dwell Rate
In an online post, Donaldson agrees that CTR is a bad metric. But he tries to replace that metric with alternatives like, “dwell time” and “dwell rate.” Just like CTR, these metrics are more likely to obscure the real value of online advertising campaigns rather than bring clarity and insights.
Most people neither want to click on an ad nor dwell on it. It’s great when people interact with ads but if that’s your primary campaign goal, chances are you are going to fail to create effective advertising and misevaluate your campaign performance.
Back To Basics
Serious discussion about whether banners work or not must first, examine the real business results. For example, do banners drive sales profitably, and if so, to what degree do banners drive greater profitability than alternative advertising formats. Do the benefits of banners outweigh the costs?
Banners may not be perfect but they’re certainly not a problem for marketers that know how to use them.
5 responses so far ↓
feesch // May 25, 2009 at 9:41 am |
Noah, thanks for linking to my article. I have said in numerous posts – to suggest “display” advertising is a failure is to write off every outdoor, print & TV ad – you simply cannot do this, in the same vein that a sign above a door is not required – it is in an awareness to something previously unconsidered, or that you are looking for.
The facts remain, in studying about 1.5 billion impressions so far, globally, that just shy of 10% of all display adverts are played with for around a minute. These facts DO NOT disprove display advertising, but actually give insights into consumer behavior that cannot be ignored – even for seemingly ridiculous formats like overlays.
Regards display to sales – every small business relies on yellow pages or equivalent. Ever large corporation relies on display across media. All are known & proven to generate favorability for the advertiser and drive sales. Even Comscore & Microsoft studies for McDonalds have proven online display drives ‘mortar’ sales.
“Other” formats are just other ways of highlighting a brand or contributing to a discussion, but these need to be tempered with scale. Creating a branded widget is not-necessarily scalable across media, is highly targeted & is much more difficult to measure in terms of ROI, for example.
In summary; yes banners are seen, watched & interacted with. Yes banners drive clicks (minimal), interactivity (marginally) & search results (proportionally) – and all exposures to brand messages sufficiently will in turn drive recall & purchase intent – online direct, online at affiliates and mortar sales.
Noah Robinson // May 26, 2009 at 4:02 pm |
Dean,
It sounds like we agree way more than I thought.
I’d love to continue the metrics discussion with you. Are you still doing a study or did you just finish one.
It sounds promising although the 10% interaction rate sounds a bit high. Is that for Eyeblaster ads (including rich DHTML) or is that across the standard distribution of ads. It’d be great if you could provide more information about the research.
For the sake of discussion, let’s assume that 10% is accurate.
Assuming a $5 CPM, 10% interaction rate, and 0:60 seconds interaction time, the equivalent CPM for 0:30 would be $25. That is pretty expensive prima facie but justifiable depending on the publisher reputation, context, targeting, what the interaction actually is, etc. It’s certainly cheaper than a 0:30 spot for online video.
Keeep in mind that I used the $5 CPM number as an example. If the cost is significantly higher (which would likely be the case for DHTML overlay ads) then the cost goes way up.
But if the cost is for a remnant display ad on a social network, the cost would be a fraction.
I think the details matter to even understand how valuable interaction is relative to other media opportunities. We’d need to dissect the weaknesses of comparative mediums (TV, print, radio) so that all formats are fairly compared against eachother.
The real 800-pound gorilla in the room is actually the impact of the interactive ads. If we aren’t comparing the actual sales and profit contribution impact of the ad (preferably using 9-digit zipcode data) then interaction rate or dwell rate isn’t that helpful.
On a side note, I think we need to start thinking about ads in terms of their square pixel size and display attention time for ads when they are actually visible on the user’s browser (n/i interaction time).
After we have that data, the next metrics might include interaction rate and CTR. But, in my opinion, neither those metrics, nor interaction rate or CTR, are meaningful without the sales data.
What do you think?
The Limits of Banner Ads // May 27, 2009 at 3:15 pm |
[...] May 24, 2009 · No Comments I’ve written before attempting to defend banner ads. [...]
CTR Is A Bad Metric For Most Brand Campaigns « Strategy 403 // June 8, 2009 at 12:57 pm |
[...] addressed this in a previous post, In Defense of Banners, but I thought it’d be helpful to expand in a single, continuously updated, [...]
Should Publishers Cut Banners « Strategy 403 // June 8, 2009 at 1:05 pm |
[...] may have issues but let’s separate fact from fiction, – Banners really, trully, actually work. – CTR is not remotely a proper measure of campaign success. – A 0.5% CTR wouldn’t be simply [...]
You must be logged in to post a comment.