Flashing text in online ads: effective or not so much?

May 26, 2007 1:33 PM

Please help a friend prove quantitatively that flashing text in online ads is a) counterproductive, and b) against best practices. Do you have any case studies, reports or first-hand experiences that may help him prove his point? Bonus points for including authorities he can cite; the ability to cite some esteemed person or institution will help strengthen his argument.
posted by necessitas to Computers & Internet (24 answers total)

Personally, I find flashing text very annoying and will generally block an ad provider that uses it.
posted by jjb at 1:40 PM on May 26, 2007

The garbage flogged via that method should be a tip-off that it's only a useful method if he's hawking a penis enlargement system, or a product pitched at the sort of adolescent girls who like, on MySpace pages, to repeatedly view their own name in animated "glitter."
posted by kmennie at 1:41 PM on May 26, 2007

The Top Ten Web Design Mistakes of 1999, by Jakob Nielsen. Number ten. (He does a collumn on web usability.) It's old, but this is an old problem.
posted by Many bubbles at 2:03 PM on May 26, 2007

Wait, here's a more recent version. #7.
posted by Many bubbles at 2:04 PM on May 26, 2007

A data point: Blinking is very effective at getting attention, but also effective is being a loud asshole. Getting attention is not the same as getting good attention. I personally feel insulted and irritated when something screams to get my attention and then I find out that it was selfishly done and isn't of at least as much value to me as the screaming meant.

I actively dissuaded two friends from purchasing home-automation equipment, because the company employed blinking ads that exceeded my annoyance threshold.

This isn't 1998. Blinking isn't novel. We've collectively learned to mentally (or programmatically) block the message. Simple and elegant is the way to go to get my attention. Charm your way in. Don't ring the doorbell every second.
posted by cmiller at 2:09 PM on May 26, 2007

Response by poster: Thanks for the response so far. Just to clarify, my friend needs no convincing that flashing text is tacky and evil. He needs to be make his case to someone who is equally convinced that flashing text is effective. In order to make his argument, he needs to be able to cite case studies, reports, etc. He's done some research, but what he's found so far has been mainly out-of-date or subjective. So, if anyone has links to these things, please include them!
posted by necessitas at 2:13 PM on May 26, 2007

There is no way to answer this question. Blinking will make the text more noticeable, but it will also make it more irritating.

The only way to decide is to look at the business model, and figure out a suitable position on the class vs. crass axis.
posted by Chuckles at 2:14 PM on May 26, 2007

I'm sorta in the biz, and there's one very elegant solution for any such (more or less) subjective quandry like this.
Officially, it's called A-B testing. It can be a little bit of a hassle, but it's the only way to *prove* stuff like this.
The idea is to create 2 ads that are identical, except for the one variable (in this case, the blinking text). Then run both ads in as equal a way as possible, and track which is getting more click-thru.
Running both ads in as equal a way as possible can be somewhat tricky of course...There are a few ways to do it though (with varying degrees of "equal").
You can run one ad for one week, then the other for the next week. (This can be a bit of an issue, if traffic to your target sites happens to be much busier on one of those weeks, but hopefully there's not *too* much variance).
If the sites they place their advertising on run a bunch of ads in the same page location (i.e. the site serves up 1 of x every time someone arrives at the page, or refreshes the site) then your friend's company could purchase 2 of the ad slots, so that both the flashing and non-flashing ad appear on the page an equal number of times.
Then, just have each ad point to a unique landing page on your friend's website (which would then automatically re-direct visitors to the actual page they want them to land on) and just track traffic to each of the unique landing pages.


A-B testing takes a little bit of time, and a little bit of hassle, but it's a great tool for turning subjective ideas like this into objective, measurable reality. And you only have to do it once, for any variable. If the non-blinking ad performs measurably better in this case, your friend will have proved their point, and won't have to have this argument next time. Of course, the blinking text might just perform better....
posted by Ziggurat at 2:26 PM on May 26, 2007 [1 favorite]

2nd'ing Ziggurat. You need a controlled test that allows you to see which is more effective for your particular case, and in this case it is an add that the only difference is the blinking text.
posted by qwip at 2:31 PM on May 26, 2007

Google's AdSense ads are unobtrusive, text-only, and extremely effective. Yes, the ads are related to your query, but it proves that ads which aren't annoying can generate business.
posted by HotPatatta at 2:44 PM on May 26, 2007

You need a controlled test that allows you to see which is more effective for your particular case

As a thought, any business looking to start an ad campaign would want to do the same thing. You could probably build on their research, either directly (by talking to someone who's done such research) or indirectly (by analyzing trends in the frequency of use of blinking ads: if it's declining, it almost certainly means it's not as profitable). Maybe not as effective as an A-B test, but it could save you a lot of effort to build off of someone else's work.

Didn't Firefox/Mozilla block the <blink> tag for the longest time? Maybe they still do? That might also be indirect proof.

He needs to be make his case to someone who is equally convinced that flashing text is effective.

Why not get your friend to convince the blinking-fan to do what Ziggurat says: in a controlled way, mix in some non-blinking ads, and see which results in more clicks / purchases / whatever exactly they're measuring?
posted by fogster at 2:48 PM on May 26, 2007

Here's the problem with A/B testing in a situation like this, as I see it. If you annoy the user, the user may stay annoyed at your campaign, even at less annoying ads. In other words, suppose I see the annoying ad. I don't click on it, because I'm annoyed, but I remember the product and the general feel of the ad. A few minutes later, on the same site, I see the non-annoying version. I don't click on that either, because the annoying ad already pissed me off. In fact, I may have already left the site and not even given myself the chance to see the non-annoying ad. Does a large enough sample size correct for that sort of thing? And if so, how do you know how large of a sample size you need?
posted by bingo at 3:45 PM on May 26, 2007

In general, people ignore banner ads. A few people even notice them, even fewer read them, and an even smaller number care about the product or service enough to follow them. It would be rare, I think, for even the most annoying ad, seen once or twice, to leave a negative, long-lasting impression on a user. Sure, a very few people might notice the ad and dislike it so much they would remember a negative impresison of the advertiser, but that would be a tiny percentage, I would think.
So yes, bingo's point isn't wrong...but I don't think it would be worth worrying about *too* much.
That does, however, raise another issue about A-B testing (and online advertising and branding in general)...
An ad doesn't have to be clicked on to "work". Even if it just raises (or lowers) the brand profile...or changes the percentage chance a user would click on an ad for the same company, in the future...it's still having an effect on those who view it. And some of that effect is not measurable through A-B testing.
So even if that blinking text surprises us all, and results in a higher click-thru than non-blinking, it *might* still be making people hate the brand.

Advertising...it's a strange and horrible art form...

But boy, all this thinking sure has made me thirsty for a Coke/Pepsi/Budweiser/Coors Light/Robitussin-on-the-rocks!
posted by Ziggurat at 4:03 PM on May 26, 2007

Looking at the site that Many bubbles recommended, I found even better proof: the results of a survey that state that 87% of internet users react to blinking or flashing ads "negatively or very negatively".
posted by The Esteemed Doctor Bunsen Honeydew at 4:03 PM on May 26, 2007

Jakob is kind of annoying in many respects ("all links must be blue, now and forever, amen"), but he gets a lot of the basics right, and he is most certainly an authority.

People have done studies that track users' eyes while they look at a Web page, and they have determined that users generally ignore the traditional top-of-page placement of banner ads entirely. Here's Jakob's story on "heatmaps".

Here's a post from David Pogue, the New York Times' respected tech columnist, arguing against blinking ads on grounds of readability.

As a person working in the field, I have found that text-based ads or highly-targeted, unobtrusive ads are far more effective than things that distract users from what they're trying to do. In terms of conversion, at least, which is the name of the game.

Good luck, and fight the good fight.
posted by lackutrol at 4:41 PM on May 26, 2007

Oh, and as someone who has worked in the field for a long time, blinking ads or "surveys" that pop up to obscure content are almost universally hated. The way to do online marketing these days is first to *make a good product* and second, to deal with your customers and potential customers honestly and pleasantly. That's how to build a good company, brand, and campaign.
posted by lackutrol at 4:47 PM on May 26, 2007

The Wikipedia page on Banner Blindness contains a number of links to the original research studies that identified the issue.
posted by ciocarlia at 5:33 PM on May 26, 2007 [1 favorite]

You aren't going to be able to prove this quantitatively.
posted by rhizome at 5:48 PM on May 26, 2007

I read a case study -- alas, I can't find it now, but it was in a small printed book that I bought from a usability organization -- that mentioned that web site users would actually _cover blinking things with their hands_ so that they wouldn't have to look at it. These might have been animations, though.

Re: Ziggurat's A/B test idea: even if clients _do_ click through, the blinking may make the company seem less solid/stable/reputable, and deter potential customers from making actual purchases. "Hey, that looks cool! I'll click it, why not... yeah, it's OK, but I'm not sure I actually want to give these people my credit card information, or my e-mail address..."

I realize this isn't the hard data you were asking for, but hey, maybe it will help.

That blinking text might do long-term, insidious damage to the company's image.
posted by amtho at 6:00 PM on May 26, 2007

Jakob is kind of annoying in many respects ("all links must be blue, now and forever, amen")

True. I do wish all websites that specify colors for links would have them change color once clicked, though. Ahem.


Oh, and as one more data point, I hardly ever see blinking, flashing, or animated ads, or ads that change when you mouse over them, because I block them (with Opera's built-in capability, and with Adblock and Flashblock when I'm using Firefox) agressively and ruthlessly and anything that's at all distracting is the first to go. (Though, if it's at the very top of the page, I tend not to bother--I just scroll down slightly to remove it from the screen.) Ads that are in the middle of text (You know--"scroll down for the rest of the article") are the second to go. Next are ads that squish text into a narrow column, or look like they're doing that.

Basically, if it's intrusive, I'll only see it once. But text ads on the edges of the page usually get left alone, because they don't scream, "NUKE ME FROM ORBIT".
posted by Many bubbles at 6:21 PM on May 26, 2007

Re: best practices. How about Section 508 of the Rehabilitation Act, the de facto accessibility standard? For a more universal standard, see the WAI/WCAG.

Anecdotally, I'm a photosensitive epileptic and flashing text makes me damned sick. I will close a website or scroll an ad off the page if it's blinking.

Also, blinking anything is just poor practice, as someone who works with usability for a living. I would never inflict it on my customers. Probably because one of my main requirements is usually "don't give the product manager a seizure."
posted by wildeepdotorg at 6:36 PM on May 26, 2007

wikipedia: "The inventor of the blink tag, Lou Montulli, has said repeatedly in interviews that he considers "the blink tag to be the worst thing I've ever done for the Internet"."
posted by Xere at 6:44 PM on May 26, 2007

It isn't clear cut, so you can't prove a point. Flashing text could mean one of those awful fit-inducing strobe light banners, or pulsing the key message once or twice to draw a little attention to it.

They need to get the discussion away from 'flashing text: good/bad?' and towards an informed, logical consideration of what will make a good ad for your particular product in that particular context. I've had various situations where ill-informed people have become crazily fixated on particular issues/'rules'; sometimes you have to patiently try to educate them, but sometimes it pays to be quite forceful and insist on setting the issue aside ("I know you're concerned about that, but the designers deal with these things every day and they'll take it all into consideration. What I'd really like to get from you is your expertise about the business/product/audience...").
posted by malevolent at 1:13 AM on May 27, 2007

HeadOn. Apply directly to the forehead.

The sad truth is that being annoying does annoy, but does work.
posted by meehawl at 11:06 AM on May 27, 2007