Part of the fun of working in the online marketing industry is watching marketers try to navigate new territory. In the dozen years or so since the commercial explosion of the web, marketers have learned chiefly by experimentation. Along the way, experience has taught marketers not only what to do, but also what to avoid.
If we're lucky, we learn a valuable lesson without spending too much money.
If we're unlucky, sometimes that lesson is learned expensively. In full view of the public. With thousands of internet users pointing and laughing.
Some of the lessons learned here are personal experiences. Others are very public learning experiences like I described above. In any case, they're clear examples of marketing blunders you need to steer clear of.
1.Failure to testNever -- never! -- going to market without testing the user experience first.
One morning in September 1997, visitors to the home page of GeoCities were
accosted by a little girl screaming, "Hey! Let me in!" at them. Not only was this seen as somewhat rude, but the problem was also compounded when many users couldn't figure out where the heck the voice was coming from.
The audio was supposed to be associated with an AT&T ad, which contained animation that would have made the whole thing make sense. But the audio produced a "WTF?" experience on GeoCities, owing to the fact that it started playing before people had a chance to scroll down below the fold and actually see the companion ad. Needless to say, Modem Media, AT&T's agency at the time, made some changes to the campaign to create a better user experience.
Though nearly a decade old, this classic blunder serves as a reminder that when we're breaking new ground with campaigns, we ought to test ads before going out to the general market to make sure the user experience is consistent with what was intended. Sometimes, both agency and client get so used to seeing an ad or campaign in a controlled test environment that they really need fresh eyes to take a look, to make sure people interpret the campaign the way they're supposed to. That fresh pair of eyes can be a formal focus group, a usability testing lab or even just a buddy looking over your shoulder. Any of these might have revealed AT&T's gaffe before it went live on some of the most trafficked sites on the internet.
2.Forgetting bandwidthYears ago, when I worked at another agency, we had a direct response client who used our agency for media planning and buying services only. A few days after launching a campaign, I sent the client some preliminary results that showed rather dismal conversion rates across the board. He called me back in a panic.
I explained that I had suspicions about his hosting provider and that he should consider upgrading his bandwidth now that his company was rolling out a campaign. The client replied that there was no way latency was responsible for the low conversion rates.
Later in the day, after I had pulled some data that showed a huge across-the-board dropoff from clicks to page landings, and supplemented that with some uptime graphs from
Netcraft, his tune changed. We paused the campaign, changed providers and re-launched. Conversion rates blew away the results we had seen previously.
In retrospect, a bandwidth check should be standard operating procedure before launching a campaign of significant size. We know about the classic examples of how live events like Victoria's Secret lingerie shows and IBM's human versus computer chess matches can bring web servers to their knees, but sometimes we forget that a large ad campaign can have the same effect.
Before launching any sort of outbound campaign that's expected to bring in a significant amount of web traffic, check with your hosting provider or IT professional to make sure the incremental demand won't slow your site down to a crawl or cause a fatal web server crash.
3.OvertargetingHere's a sin of which many online marketers are currently guilty. The advanced targetability of online advertising is simultaneously a huge benefit and a major drag.
The benefit, of course, is that advertisers can use targeting to concentrate their media efforts on the sweet spot of their target audience.
The drag is that a lot of advertisers confuse "sweet spot" with "entire market," and they overtarget their ads, often with detrimental results.
A quick war story: in 1998, I worked on a telecom account where the client was advertising ISDN and DSL within a certain geographic footprint. We employed geotargeting in order to minimize waste and to cut back on the number of folks ringing up the call center from outside the service area. Along came a proposal from a well-known search engine to serve the client's advertising only to people looking for ISDN or DSL within the target geographies. Geotargeted keywords sound like a good idea, right?
Wrong. As it turned out, the search engine had location data on only a small fraction of its user base. Overlaying the audience of folks searching relevant keywords with the tiny pool of folks on which they had geographic data resulted in a microscopic potential audience, indeed. After slamming the same 11 people over the head a few hundred times with our ad, we canceled the order.
In retrospect, we shouldn't have overlaid the two targeting filters. ISDN and DSL had such low penetration compared to dial-up back then that we could have easily forgotten about the keyword filter entirely and had a successful campaign.
These days, advertisers make the same mistake very frequently. If the demographic sweet spot for red rubber balls is Men 18 to 24, online campaigns are often focused on that demo with laser-like precision. And that's a shame, because there's really nothing preventing a 35-year-old woman from buying a red rubber ball.
Don't assume a target audience is a market, or vice versa. One of the differences between television and interactive is TV's spill into demographic audiences other than the one targeted. While certain television campaigns might be guaranteed against Men 18 to 24, they also deliver some Women 25 to 49 in addition to the buying target. Some of those women buy the product. Why should it be different for online ads?
Experiment with loosening your targeting restrictions a bit. You might be pleasantly surprised by the result. Obviously, your mileage may vary.
4.Failing to cap frequencyAccording to Atlas Solutions, advertisers can decrease their cost per acquisition on direct response campaigns, simply by setting an appropriate frequency cap. The trick is setting the cap at the appropriate level. Direct marketers have always had to play the trade-off game between efficiency and volume, and while a frequency cap of 3x might generate conversions at a lower Cost Per Acquisition, a frequency cap of 5x might result in a much higher volume of conversions at a CPA that still meets the advertiser’s goals.
Without naming names, we all know advertisers who have become infamous for saturating the Internet with ads that we end up seeing over and over.
Audience chafe factor might not be a good enough reason for some advertisers to consider capping frequency. After all, if given the choice between getting the order as the result of being in a prospect's face 24/7 and not getting the order because you weren't top of mind when the prospect went shopping, most direct marketers would choose the former. But most studies of both direct response and branding campaigns
uncover a distinct point of diminishing returns. For DR campaigns, there's a sizeable falloff after the third impression, and optimal frequency (with respect to profitability) comes just a few impressions later. For branding campaigns, optimal frequency tends to vary, but most advertisers can quickly discern the point of diminishing returns by glancing quickly at their last Dynamic Logic brand study (or equivalent).
Once that point of diminishing returns is reached for an advertiser, doesn't it make sense that continuing to bash people over the head with an ad is a giant waste of money? Before you answer, consider that frequency-capping offers you the opportunity to stop advertising to the inundated and serve those impressions instead to people who haven't seen your ad yet or who are underexposed to it.
If you pay for your ads on a CPM basis, the only ads you're running that should be exempt from frequency-capping are sponsorships where people expect to see your ad every time in a given location. Everything else should be optimized, regardless of the goal, based on optimal frequency.
5.
Affiliate carte blanche
"Get me 10,000 new customers by the end of the quarter. No, I don't care how," is an oft-repeated mantra, especially for direct response advertisers. What many advertisers don't understand is that this is riskier than tap dancing on a land mine.
With carte blanche to do whatever it takes to generate customers, an affiliate can and often will use all sorts of unsavory tactics, including creating its own affiliate program comprising all sorts of sketchy folks who will also use any and all means at their disposal. Those means are often along the lines of spam, phishing scams, bogus co-registration, spyware and all sorts of other nasty (and likely fraudulent) things.
Ever see a piece of spam come into your inbox from an otherwise respectable brand and wonder why the marketer would resort to spam? This is precisely how that happens. Someone on the marketing team needs to reel in an affiliate.
In the recent past, an affiliate of one of our clients was tasked with generating a certain number of registrants for the client. When they found they couldn't live up to their commitment, they moved from respectable marketing tactics toward shady ones. When the registrants they referred immediately opted out from my client's offering upon receiving email confirmation, it was a huge red flag that something had gone awry. In this case, the affiliate had simply -- and shadily -- force-subscribed people in its own database to the client's offer. Thankfully, we had guidelines in place such that the client could immediately sever all ties with the affiliate without penalty.
Such guidelines are a necessity if advertisers are going to do business in the affiliate marketplace or in the realm of CPA advertising. Otherwise, the brand is at risk of both legal exposure and being associated with shady marketing tactics.
6.Managing paid search by handIt's one thing to plunk down $50 a day for a few keywords if you have a small business or simply want to test out paid search. It's another thing entirely to manage your bids on a sizeable paid search campaign manually over a period of time.
A competitor of one of our clients once confessed that we were driving his team of on-site search specialists positively bonkers. The team would update its bids, hoping to "jam" my client by underbidding us by a penny, thus increasing our click costs. Mere minutes later, we would counter-jam his team, forcing them to go back and re-bid everything. What the competitor didn't realize at the time was that he was bidding against a system, not human beings, and that our bidding strategy was based on a collection of logic rules-- some simple, some complex.
A good bid management system (
Atlas Search,
24/7 Search,
DART Search, et cetera) pays for itself almost immediately by increasing your bang for the buck and keeping your paid search costs down. You'll also become a lot more efficient through cutting back on the manpower that you've allocated to managing your search campaign.
Given all this, I'm flabbergasted when I continue to run into advertisers who insist on managing bids manually. All it takes is a single competitor's employment of a decent bid management and search tracking solution and any manual management you might be doing will be rendered completely ineffective for terms you compete on. Much like a martial artist uses jiu jitsu techniques to apply your own energies against you, a competitor will use a bid management system to increase your click costs, take all the good inventory for themselves, and blow you out of the water from an efficiency standpoint.
Investing more than $1,000 a month in paid search? Don't do it manually.
7.Ignoring mindsetI've already mentioned testing the user experience before launching a campaign to make sure the experience is consistent with the expectations of the advertiser. Another user experience sanity check that needs to be conducted before launching an online campaign is the mindset check.
A client of ours once tried to run an ad that took nearly 30 seconds to run through its animation sequence on a site where the typical user was in and out in half that time. Obviously, folks didn't stick around long enough to see that ad play out. But that's just one symptom of an overarching disease.
The sanity check has to do with the mindset your prospects will tend to be in when they encounter your ad. Are they browsing? Are they "mission critical?"-- that is, are they preoccupied with hurriedly completing a specific task? Are they on a page that typically serves as a quick gateway to another?
Ignore mindset at your peril. Many advertisers filter inventory from their buys based on mindset. They might not buy ads on Instant Messenger applications, knowing that people are less likely to respond if they're actively engaged in a chat with a buddy. They also might ignore proposals to buy ads on stock quote pages, knowing that even the desirable individual investor target can be less attentive if they're quickly typing in one stock quote after another as they check on prospective investments.
Creative can be optimized according to mindset. One might use a short-form animated banner on a gateway page and a longer form creative on a page where people are likely to spend more time. But mindset shouldn't be ignored. Doing so can reduce the effectiveness of your campaigns and cause perfectly good inventory to fail to produce brand impact and increased response.
8.Not customizing sponsorshipsSo you bought a terrific sponsorship that involves three ad units on a page. The leaderboard on the top, a nice skyscraper down the right-hand side and the leaderboard on the bottom.
Now is not the time to run the same ads you run in ROS rotations.
You can see these sponsorships all over the web, where the advertiser owns the page and every ad unit on it. Such multi-unit sponsorships have a terrific chance of making a big impression on a potential customer, but all too often advertisers fail to take advantage of the space they've purchased.
Some of the best sponsorships use multiple ad units in concert. Think Budweiser and its multi-unit ad that poured beer from the top leaderboard into a frosty glass waiting in the side skyscraper. The worst sponsorships simply fill the space with three random ads rotating elsewhere on the buy, and the user ends up inundated with different messages for the same product that compete for attention.
When you buy interesting space, you need to fill it with interesting advertising that takes advantage of what you bought.
9.Forcing subscriptionsSo you've got a giant database full of people who have purchased red rubber balls from your company, and you want to start a rubber ball newsletter. It would be so easy to simply export that customer database into the seed list for your new newsletter, wouldn't it?
As Yoda once said, "Once you start down the dark path, forever will it dominate your destiny."
As easy as it would be to give your newsletter a running start, force-subscribing customers to anything is a great way to get them to run away from your company, screaming.
I'm not kidding about the screaming part. People love to talk in online forums about how Company X spammed them and how they'll never do business with them again.
Don't forget, when someone spouts off about being spammed on a blog, message board, social networking site or other online community, those comments tend to stick around. And we all know how well blogs and community sites tend to be treated by Google and other search engines. A negative comment about spam out in the blogosphere, especially if other bloggers link to it, can be devastating to a brand's online reputation, not to mention natural search engine results.
Like it or not, a newsletter needs to be built organically. Perform an audit of customer touchpoints, from your website to your advertising to even your email signatures. You'll find plenty of opportunities to get customers and prospects signed up through an opt-in process.
10.Acquisition vs. retentionThe ability to separate your existing customers from your prospects is one of the things that make traditional database marketing very effective. After all, it's a lot easier to upsell or cross-sell an existing customer than it is to pluck an entirely new customer from the universe of folks who have never bought anything from your company.
There's no lack of opportunity to personalize messaging to brand loyalists, switchers and new prospects online. Behavioral targeting, remessaging, profile targeting and email personalization are all tactics we can use to segment and personalize.
Why, then, do I see a good number of campaigns that treat existing customers like they've never interacted with the advertiser before? In many cases, this can be downright insulting.
By way of example, a few weeks after making a big-ticket purchase with an online retailer, I received the same acquisition message in an email that any prospect might get. The email contained a pitch for one of the things I had just bought from the company.
The fact that this happens in a dynamic marketplace where personalization is easy says something to me. It says the retailer doesn't care enough about my business even to acknowledge that I've been a loyal customer.
How difficult would it have been for the retailer to look up my purchase history, insert a line in the email that says, "We appreciate your business," and recommend a product for me that wasn't the one I just purchased?
Not very difficult at all.
Fact is, CRM solutions make it easy for marketers to mirror online what they're doing in traditional database marketing. Ad servers can easily help to differentiate between folks you've already reached with your ad campaign and folks who are seeing your ads for the first time. They can differentiate between different types of customers, too. So why deliver the same ad you run to attract new customers to the folks who already buy from you?
Conclusion: You live, you learnNo one expects marketing to be perfect 100 percent of the time. That's why we test a lot in this business. But the expectation is that once we've learned from our testing, we should be applying that knowledge wherever we can to improve our marketing efforts continually.
Each of the 10 marketing blunders I've described here is a lesson someone learned, sometimes the hard way.
The good news is that their mistakes are your learning opportunities. Don't doom yourself to repeating history.
Last suggestion: print this piece out and see if your company is currently committing any of these blunders.
If so, fix them.
MEDIA PLANNING & BUYING: IN FOCUS
Apple often gets credit for starting the personal computer revolution, but the Macintosh, which debuted in 1984, was not the original mass-market PC. On Aug. 12, 1981, 
