After Wave of Disappointments, The Web Lures Back Advertisers
Mitsubishi Motors Corp. is hurting from plummeting U.S. sales, and its North American unit is cutting costs drastically. Since October, it has halted plans to expand its Illinois factory, announced hundreds of U.S. layoffs and restructured its executive ranks.
But the Japanese auto maker is boosting spending in a surprising area: ads on the Web. Mitsubishi says it plans to increase online-ad spending as much as 50% this year, to $6 million, as part of a larger marketing push.
Just three years ago, Mitsubishi abandoned online advertising as ineffectual. Now, says Carlos McEwan, who oversees Mitsubishi's Web ads, "the Internet is a permanent fixture for us."
Web ads, not long ago a wasteland of over-hyped expectations, are making a comeback. After rushing to advertise online in the 1990s, many companies wound up disappointed and dropped Internet ads altogether. Today, some are coming back, lured by an improving economy, lower rates and better technologies for tracking potential customers.
Online ad spending climbed 20% last year to $7.2 billion. Yahoo Inc.'s ad revenue surged 84% last year. Analysts estimate revenues were up threefold at the closely held search titan Google, which doesn't disclose results. Forrester Research Inc., a market-research firm, estimates Web-ad spending will grow as much as 23% in 2004.
Many advertisers have figured out how to use techniques that only the Web can offer -- including ways that let them capture names and data of consumers, measure results of Internet-ad campaigns and track the paths of consumers from the ads they view to the date and place of purchase.
DaimlerChrysler AG increased its online media spending by 30% last year. Toyota Motor Corp. increased its ad spending 55% in the first nine months of 2003, compared with all of 2002. Delta Air Lines has raised online-ad spending. Huffy Sports, the sports-equipment unit of Huffy Corp., cut back sharply on Web ads in 2000, but is now increasing such spending because of better tracking capabilities. The Web delivers something TV, radio and billboards can't, says Brian Meehana, Huffy Sports spokesman. "You can build a relationship, ask them questions, have a back-and-forth" with consumers. "It's the only medium you can do that in."
The growth figures are coming off a shrunken base after the bursting of the late 1990s dot-com bubble. Many big advertisers willing to spend limited amounts on an online product campaign aren't making longer-term commitments. And the Internet is widely seen as a good way to reach targets such as teen-agers or young adults, but not a mass audience.
Ken Burkeen, assistant brand manager at Procter & Gamble Co. says Web advertising is still being held back by the difficulty of reaching a large audience all in one place, and the lack of an accepted way to measure the effectiveness of the medium. "There is no industry standard," he says.
At Mitsubishi, Web ads this year will account for only about 5% of its overall advertising budget. But the company is going back online because it thinks it has figured out how to guide Web surfers to showrooms. Mitsubishi says that it has found online ads are more effective than other pitches in leading buyers to dealers.
Web ads helped Mitsubishi get Tom Buczkowski to Dennis Beres' showroom. Last year, Mr. Buczkowski, a Silver Creek, N.Y., police officer, was in the market for a sport-utility vehicle. He spotted an ad on the car buyers' Web site Autobytel: Sign up for a test drive, the ad offered, and win a $15 gas card. He clicked on the ad. It relayed him to Mitsubishi's Web site. There, the 56-year-old Mr. Buczkowski "built" an online model of an Outlander SUV outfitted to his tastes.
Then Mr. Buczkowski did what most Web advertisers dearly want consumers to do: He typed in his name, address and phone number. That entered him in a sweepstakes for a free SUV, and signed him up for the gas card. It also sent his data, in minutes, to Mr. Beres, a salesman at nearby West-Herr Mitsubishi in Hamburg, N.Y. The site asks customers to type in their data if they want to hear more about a vehicle, then sends the information to the dealership in the nearest ZIP code.
The next day, the dealer phoned Mr. Buczkowski to urge him to take a test drive. He was impressed: "If you get a response from somebody a day after the request, that's something you don't find too often anywhere else." He took the test drive two days later and eventually bought a Mitsubishi in November.
Such ads differ from the Web pitches of just a few years ago, which did little to exploit the Internet's capabilities. Like many companies, Mitsubishi rushed to spend on Web ads in the late 1990s. Among its early efforts: big, expensive "banner" ads, that were little more than online versions of print ads. In Mitsubishi's case, they led Web surfers to pages with pictures of cars that were the Internet equivalent of the car maker's print brochures.
Like most Web advertisers during the dot-com boom, Mitsubishi tracked "impressions," or the number of times that people viewed a page. It also tracked "click-through rates," the number of computer users who clicked on its ads. Mitsubishi, like other advertisers, wasn't quite sure what those measurements meant or what it wanted out of online advertising. "It was more the promise that was exciting, not the actual results," recalls Mr. McEwan, who participated in Mitsubishi's marketing meetings.
It didn't seem to matter. Mitsubishi's rivals were all spending online with abandon, as were many U.S. consumer companies. Overall Internet advertising spending peaked at $8.1 billion in 2000.
But as the dot-com boom wore on, Mitsubishi executives began to wonder whether Web ads worked. Greg Stahl, Mitsubishi's former ad director, questioned the value of online banner ads, which he personally regards as "a nuisance." Mr. Stahl and his team began examining the ads' impact more carefully. There was no evidence, they concluded, that Web ads accomplished their key mission: leading consumers to showrooms.
In August 2001, Mitsubishi halted all new spending on Web advertising. "The thinking was let's not send a lot of people to our Web site if we're not sending any leads to our dealers," says Mr. McEwan.
The withdrawal from the Web mirrored similar decisions by hundreds of other advertisers. Deciding that Web ads didn't produce measurable results and pressured by the economic downturn to be more selective with advertising dollars, corporations pulled their online-ad budgets. Conventional wisdom at the time was that online advertising didn't work -- and might never. Yahoo and America Online, which once turned away ads, began to feature more "house ads," or ads from the site itself. Yahoo saw a 35% drop in online ad revenue in 2001 from 2000. Online ads overall dropped 12%, and then 15% the next year.
While Mitsubishi stopped buying online ads, it kept looking for ways to exploit the Web. One tip came from a Texas dealer, who was already alerting his salesmen about prospects that came in from the dealer's own Web site. His finding: The sooner a dealer responded to a Web lead, the better the chances were of selling a car.
For online advertising to ever work, Mitsubishi concluded, it had to quickly get the Web surfer who clicked on an ad in contact with a car dealer. The company's programmers reworked its Web site to relay Web visitors' contact information almost instantly by e-mail to car dealers. Mitsubishi began training dealers to check their e-mail boxes often for the leads and quickly call prospective buyers.
Meanwhile, the broader ad downturn was spurring the industry to solve some problems that had helped drive advertisers away. Advertising-technology companies such as DoubleClick Inc. and Tacoda Systems built technologies to track Web users' activities and then deliver targeted ads to those visitors. Advertisers were discovering "search-related ads" -- text ads that appear when a person conducts a Web search. In addition, prices were falling: Published ad rates charged by Web site operators plunged 23%, according to market research firm Nielsen/NetRatings.
Still, Mitsubishi didn't have high hopes when, in July 2002, it dipped a toe back into Internet advertising with a two-month campaign on Yahoo and a few other sites to pitch its Outlander SUV. Past Internet ads did little to get shoppers to hand over personal information. The new ads encouraged Web surfers to test-drive an SUV by entering a sweepstakes or applying for free gas cards. The site was designed to quickly send personal information to Mitsubishi's ad agency, which then relayed it to dealerships.
The campaign generated more leads than the company expected. Mitsubishi found that the campaign reduced its cost per sale -- the amount of money spent on advertising divided by the number of cars sold -- to one third of the cost of traditional advertising media, says Mr. McEwan.
Around that time, Internet ad spending began turning around, reaching $1.7 billion in the second quarter of 2003, the third consecutive quarter of growth. Traditional advertisers started noticing that customers spent substantial time on the Web for research on topics such as home loans and travel packages, as well as e-mail and shopping.
Mitsubishi says it was generating better-than-expected results when it ran online promotions around the introduction of different vehicles, which impressed top managers. "They looked at it and said, 'This thing is working,' " Mr. McEwan says. The company increased spending after each successful Internet campaign by slicing money out of its billboard and print publication budgets. When it advertised on the Web, it says it saw a tangible, more immediate impact on sales.
Advertisers were finding that Web-site operators were also being more responsive and finding ways to fix campaigns that weren't working -- a big change from the dot-com days, when sites offered banner ads, take it or leave it. And advertisers have been experimenting with lots of new things since the dot-com bubble burst: bigger-sized ads, animation that allows ads to flit across the Web page, use of video and more creative "image" ads that help boost retailers' brands among computer users.
In the spring and summer last year, Mitsubishi experimented with "spot buys" on the Web, short-term Internet ads that were bought last minute, to promote customer-service special offers such as extended vehicle warranties and summer clearance events. Response was lackluster. "The big lesson is you can't run something for a couple weeks and expect it to generate much," said Mr. McEwan.
Mitsubishi continues to tweak its Web ads to shrink the time between consumers' clicking on an ad and entering a showroom. Before November 2003, Web surfers who visited the Carsdirect.com Web site would see an ad for Mitsubishi's Galant sedan. The ad offered a test drive and a chance for a $25,000 shopping spree. Click on the ad, and a Web surfer would land on a page detailing the car's features. Another click and consumers would land on a questionnaire page, which then relayed personal data to Carsdirect's computers.
That was too slow, decided Mitsubishi's ad agency, 10th Degree: The clicking consumers needed to be asked immediately for personal information before they had a chance to think twice. And that data needed to go instantly to Mitsubishi's servers, instead of being delivered in batches via Carsdirect's servers. Carsdirect quickly altered its data-transfer process; electronic leads from Web surfers now got to dealers in 10 seconds, down from five minutes.
Mitsubishi also tried "behavior-based advertising," a technology offered by Yahoo and others. The technology monitors computer users' surfing habits and displays Mitsubishi ads to people who have visited the Yahoo Autos Web site or searched for car-related information by typing in the phrase "buy a car." So someone who had visited the cars section on Yahoo might see a car ad when checking their e-mail or surfing the company's music site. Publishers and advertisers figure that the likelihood is high that someone who has visited car Web pages or searched for auto information is shopping for a car.
网络广告卷土重来
三菱汽车(Mitsubishi Motors Corp)眼下因美国销售额骤降而一筹莫展，其北美子公司正大刀阔斧地削减成本。自去年10月以来，三菱汽车扩建伊利诺伊州工厂的计划已经告停，而且还宣布在美国裁员数百人并对管理层进行重组。
但该公司却在一个说来非同寻常的领域扩大了支出：互联网广告。三菱汽车称，作为加大营销力度的一项举措，公司今年计划将互联网广告支出增加50%之多，达600万美元。
就在3年前，三菱汽车因网上广告收效欠佳而放弃了这一促销手段。而今负责该公司网上广告业务的卡洛斯?麦克伊万(Carlos McEwan)却表示，互联网是公司永远的营销媒介。
不久之前，网络广告还因声过其实而遭冷落，但如今它又开始恢复元气。在90年代蜂拥推崇网络广告后，很多公司最终因失望而纷纷退出。今天，鉴于经济持续复苏、利率水平较低，并且网上广告招揽客户的手段也更加高明，因此一些广告客户又回来了。
去年网上广告支出总额增长了20%，达到72亿美元。雅虎( Yahoo Inc.)去年的广告收入猛增84%。分析师估计搜索巨头Google的广告收入甚至增加两倍，但Google没有公布具体数据。市场研究公司Forrester Research Inc.预计，2004年网上广告支出将增长23%。
很多广告客户已经领悟到网络广告的独到之处，包括可以在网上获得客户的姓名和资料，可以衡量互联网广告促销活动的成效，并且可以跟踪客户的喜好，从他们对广告的评价到购买商品的日期和地点等无所不能。 去年，戴姆勒克莱斯勒(DaimlerChrysler AG)将其网上广告支出提高了30%。2003年前9个月，丰田汽车(Toyota Motor Corp.)将广告支出提高55%，相当于2002年的全年支出。达美航空(Delta Air Lines)也提高了网络广告支出。Huffy Corp.的体育设备子公司Huffy Sports在2000年大幅削减了网络广告支出，但目前考虑到网络广告具有更好的跟踪能力，该公司也开始增加这方面的支出。Huffy Sports的发言人布莱恩?米汉(Brian Meehan)说，网络广告可以传递一些电视、广播和户外广告所不能传达的内容。你可以在网上与客户建立联系，向他们提出问题，并能得到反馈。网络是唯一具备此功能的一种媒体。
然而，这种广告支出的增长数据是在较低的基数之上得出的，在90年代末网络泡沫破灭后，网上广告一度萎缩到极低的水平。愿意将有限资金用于网上营销的很多大型广告客户也并未做出长期承诺。人们广泛认为互联网是接近年轻人的良好方式，但并不是联络普通大众的首选。
宝洁公司(Procter & Gamble Co., PG, 又名：宝硷公司)品牌经理助理肯?伯基恩(Ken Burkeen)说，商家对于网上促销方式仍然畏首畏尾，因为网上广告很难同时触及到大批客户，并且目前仍缺乏一个能够被大家认可的衡量媒体效率的标准。他说，媒体业没有行业标准。
至于三菱汽车，其网上广告支出今年将近占其广告总体预算的5%左右。但该公司还是返回网络了，因为它已发现如何把网上的买家吸引到经销商那里。三菱汽车称，它发现在吸引顾客达成交易方面，网络广告比其他促销方式更为有效。
正是网络广告将56岁的汤姆?巴克佐柯斯基(Tom Buczkowski)吸引到了三菱车推销员丹尼斯?贝雷斯(Dennis Beres)所在的汽车展厅。去年，这位在纽约银溪工作的警官正想买一辆运动型多用途车。他在购车人网站Autobytel上看到这样一则广告：报名参加试驾活动吧，还可获价值15美元的汽油票。他点击广告后，进到了三菱汽车的网站。巴克左柯斯基在三菱汽车网站上看上了一种适合他口味的欧蓝德(Outlander)运动型多用途车。
然后，就像大多数网上广告客户最希望的那样，作为潜在消费者的巴克左柯斯基输入了他的名字，然后是地址和电话号码。在这番操作之后，他不仅将得到汽油票，而且如果运气好，他还有可能免费获得一辆运动型多用途车。几分钟后，网络将他输入的信息传到了贝雷斯那里。贝雷斯是West-Herr三菱汽车服务中心的销售商，该中心位于纽约州汉堡，与巴克左柯斯基的家在同一地区。如果一位消费者希望了解有关一款汽车的更多情况，网络会要求他输入相关数据，然后网络再将这些数据转给与消费者所在地区最靠近的经销商。
第二天经销商就给巴克左柯斯基打电话催他去试驾了。他对这件事印象很深。他说，在你提出要求后第二天就有人给你答复了，这种事情在任何其他地方可都不常见。两天后，他去了，最后，在去年11月份，他买了一辆三菱车。
这样的网络广告与几年前相比已不可同日而语。与许多公司一样，早在90年代末，三菱汽车就很热衷在网上作广告了。早期的广告中有一种是价格昂贵的大尺寸横幅广告，这种形式的广告充其量也就相当于纸上广告的网络版。具体对三菱汽车的广告来说，网上冲浪者可以点击浏览车型，就像在网上浏览三菱汽车的广告小册子差不多。
跟"网络热"时期多数发布广告的厂商一样，三菱汽车关注的是点击率，即人们浏览某页面的次数。它还关注"点进率"，即人们点击广告图标进入后续内容的次数。但三菱汽车跟其他厂商一样，对于这些指标意味著什么以及它自己想通过网络广告达到什么样的效果心中没数。当年曾参加过三菱汽车营销会议的麦克伊万回忆说，那些点击率、点进率在更大程度上只是一种令人兴奋的前景，而不代表什么实际结果。
但这并不重要，因为三菱汽车的竞争对手们也都在网络广告上一掷千金，许多美国公司也是这样。2000年，互联网广告总支出高达81亿美元。
但随著网络热逐渐降温，三菱汽车也开始思考网络广告是否真的有效。三菱汽车前广告主管格里格?斯塔尔(Greg Stahl)对网上横幅广告的价值提出质疑，他个人认为，这类广告只会让人讨厌。斯塔尔和他手下的人员开始更细致地研究网络广告的效果。他们最后得出结论认为，没有迹象表明网络广告像厂商们希望的那样，起到将人们吸引到汽车展厅的作用。
到2001年8月，三菱汽车停止发布新的网络广告。麦克伊万说，他们的想法是，如果（网络广告）不能将很多人送进经销商的销售展厅，也就没有必要将人们送进三菱汽车的网站看广告。
不独三菱汽车如此，其他数百家发布网络广告的厂商也作出了类似的决定。鉴于网络广告不能产生任何切实的效果，而且在当时经济下滑的情况下各家公司在决定广告支出的投向时压力很大，因此，它们纷纷削减网络广告预算。
那一时期人们普遍认为，网络广告没什么用，而且可能永远都不会有用。雅虎和美国在线(America Online)后来登载的广告更多的是所谓"内部广告"，即它们自己的广告。2001年，雅虎的网络广告收入较2000年下降了35%，而网络广告市场的总体规模下降了12%，2002年进一步下降15%。
尽管三菱汽车不再购买网上广告版面，但它一直在探索利用网络的各种方式。其灵感之一来自得克萨斯的一家三菱车经销商。以前这位经销商就提醒手下的销售人员关注他们自己的网站上收到的消费者信息。他发现，经销商对网上传来的有意购车者的试探信号反馈地越快，最后达成交易的几率就越大。
三菱汽车认为，要想让网络广告发挥作用，必须要使点击广告的人能迅速与经销商联系上。
三菱汽车网站的制作人员对网站进行了改造，以使网站访问者的信息能通过电子邮件即刻传送到经销商那里。三菱汽车已开始对经销商进行培训，要求他们经常检查邮件并快速给有意购车的人去电话。
同时，广告大幅下滑的惨痛经历已经促使网络广告行业开始解决那些导致广告客户退避三舍的问题。
DoubleClick Inc.和Tacoda Systems等广告技术公司开发了一些能跟踪网络用户浏览习惯的技术，这种技术随后还能向网络用户传送有针对性的广告。
广告客户还发现有一种针对搜索内容的广告，也就是说，网络可以在用户检索内容时弹出与之相关的文本广告。另外，网上广告的价格也在下降：据市场研究公司Nielsen/NetRatings提供的数据，网站运营商对已发布广告的收费标准已下降了23%。
然而在2002年7月，三菱汽车当时并未抱有太大希望，只是在雅虎和少数几个网站上为其欧蓝德运动型多用途车进行了为期两个月的广告宣传而已。以前的互联网广告几乎不要求购物者提交个人信息，而新的广告则鼓励网友参加抽奖或者申请免费油票来获得试驾机会。网站通过特别的设计，可以将个人信息迅速传送至三菱汽车的广告公司，然后由他们转交给三菱汽车的代理商。
这次促销的效果好于公司的预期。麦克伊万说，三菱汽车发现，这种做法的每笔交易成本仅占传统广告方式的三分之一。每笔交易成本即广告费用除以所售车数量。
大约在那个时候，互联网广告支出开始反弹，2003年第二季度达到17亿美元，实现了连续第三个季度的增长。传统广告商开始注意到，客户花费不少时间上网搜索各种信息，包括住房贷款、旅行社组团计划以及购物信息，同时也收发电子邮件。
三菱汽车表示，在推出不同车型时进行网上促销的收效比预期的更好。每一次成功的网络促销后，公司都会削减广告牌以及印刷物广告费用，转移到网络广告上来。三菱汽车认为，网络广告对销售的推动更明显，也更迅速。
广告商们发现，网络运营商开始提供更多反馈意见，想办法调整低效广告。这与互联网狂潮时相比大不一样，那会儿网络运营商只提供一种横幅广告方式，别无选择。网络泡沫破灭后，广告商们也体验了大量新的事物，运营商们提供了更大面积的广告、在页面上飞来飞去的动画广告、或者通过录像和更有创意的"形象"来提升零售商品牌在网络用户中的知名度。
去年春夏两季，三菱汽车尝试了网络"现场购买"方式，这是一种短期的互联网广告，促销公司的客户服务特别优惠项目，例如延长保修期，进行夏季轻货等。但反应寥寥。麦克伊万表示，得到的教训是不能期望一个短期广告带来大量业务。
三菱汽车继续调整其网络广告，以期缩短客户从点击广告到进入网络展厅的时间。2003年11月以前，访问Carsdirect.com网站的用户应该能看到三菱戈蓝(Galant)轿车的广告，提供试驾和2.5万美元的折让。点击广告，首先进入一个介绍戈蓝轿车详情的网页，再点击一下就进入一个调查表的页面，用户所填个人信息会转往Carsdirect的电脑。
三菱汽车的广告公司10th Degree觉得这也太慢了一些。他们的想法是，点击广告的客户首先应该按要求提供个人信息，然后在仔细考虑是否购买。另外，这些信息应该立即传送至三菱汽车的伺服器，而不是由Carsdirect的伺服器分批转发。Carsdirect迅速调整了数据传输程序，现在从客户提交数据到代理商接到数据之间仅需10秒钟，以前可得耗时5分钟。
三菱汽车也尝试了雅虎和其他几家网站提供的"行为方式广告"。这种技术可以观察用户的浏览习惯，对那些访问雅虎汽车频道，或者打入"买车"在搜寻相关信息的用户面前显示三菱汽车的广告。所以，曾经在雅虎浏览过汽车信息的用户就可能会在查收邮件，或者浏览雅虎音乐网站的时候看到一条汽车广告。网站和广告商的想法是，曾经浏览过汽车网页，或者搜寻过汽车信息的人多半是要买车的。
