Just like everything else in search marketing, paid search can be broken down into a series of steps that you can take, one by one. Paid search might look complicated, and sometimes it is. Every search engine has different program rulesand they change. You will never be able to predict what your competitors will do next, which has a huge impact on your results. By focusing on each step in the process, however, you can make paid search as simple as possible. Here are the high-level steps that every paid search program must take:
Those are a lot of steps, so let's get started with the first one, setting up your paid search marketing program. Set Up Your Paid Search ProgramSetting up your program has three major tasks: deciding your budget, setting up your accounts, and choosing your bid management software. We start with the budget. Decide Your BudgetWe spent some time in Chapter 8 developing a budget, but you might need a more precise method than we showed you there. Unfortunately, it is not all that easy to do. The tools you have at your disposal limit the amount of information you can gather, and what you ultimately spend depends in part upon what your competitors do, which is unpredictable. However, you can make better guesses about your budget. The first thing to do when setting your budget is to decide how you want to do it. Do you want to build to budget? Or budget to build? In other words, will you add up all your possible costs and ask for that budget (budget to build)? Or are you being handed a budget to do whatever you can afford under the budget (build to budget)? Either way, your biggest challenge will be to properly estimate the costs. As you recall from Chapter 3, there are three major costs in paid search:
In Chapter 8, we walked through a relatively simple process for projecting a paid placement budget for Yahoo! By examining your keyword demand in each paid placement engine for each campaign, and by religiously following the technique from Chapter 8, you can take a stab at your media costs for any paid placement engineexcept Google. Google is especially tricky to project costs for, because of their clever ranking algorithmthe one that shows your ads based on a combination of your bid rate and your clickthrough rate.As discussed earlier, Google's ranking method is a wonderful way to reward search marketers with relevant adsif you design your ad to get a higher clickthrough rate than your competitor's, you will get a higher ranking for the same bid. However, the same cleverness that can give you that edge wreaks havoc with your attempts to budget your Google AdWords campaign. Here's why. With Yahoo! (and all other paid placement engines except Google), you can take a guess at what your bid must be to get a particular ranking. You can look at the bids from the current bidders for a keyword and estimate that your bid must be in that top range to get a top rankingto be conservative, you might estimate that your bid must be significantly higher than the current #2 bid if you want to rank #2, for example. You are taking a guess because you cannot predict how bidders will change their bids in the future; they might even change their bids based on your bid, because the former #2 bidder wants to regain that spot. But with Yahoo!, you can at least get a feeling for what your bid must be to get that #2 ranking, for example. With Google, you cannot. Google will show you an "average bid," but it cannot offer any way for you to target your bid to achieve any particular ranking. It cannot because the clickthrough rate is just as important as the bid in determining the ranking. If your ad gets heavy clicking from searchers, it will take a lower bid to rank #2 than if it has a below-average clickthrough rate. Because you cannot easily project your ranking in Google, you also cannot easily project your clickthrough rate, because you know that your rate in part depends on where you are ranked. (Does your brain hurt yet?) In Yahoo!, you take your guess at your click rate and multiply it by the keyword demand (the number of searches for that keyword) and you can estimate the number of clicks you will get. Multiply the clicks by your bid, and you have a projected cost. In Google, you can only get an average number of clicks based on your average bid, so your projections could be wildly off if you bid higher or lower than average and if your clickthrough rate diverges from the average (higher or lower). (Now do you have a headache?) But there is even more complexity, unfortunately. If Google's ranking formula were as simple-minded as multiplying the click rate by the bid, no new ads would ever be shown, because their initial click rate is zero. So, Google actually has a default clickthrough rate it uses to put new ads into rotation, and waits for a certain number of impressions before applying the ad's actual clickthrough rate to determine its ranking. (There, now you are finally reaching for the Excedrin.) So what do you do? You have a few options:
Snap Electronics initially chose to use only Yahoo!, and they set a budget for their Yahoo! campaign back in Chapter 8. But Snap eventually wants to use Google, too, so they investigated what a Google budget might be. In Figure 14-3, you can see that the average digital cameras bid costs 85¢, the same price as Yahoo!'s top bid, but the way Google displays paid results makes it tougher to estimate your true costs. Sometimes the rates vary widelythe bid for digital camera (without the s) averages $1.23. Snap can use the cost per day to decide whether its estimated budget should be higher or lower than the averages. Figure 14-3. Snap's Google bid limits. Google shows both the maximum and average bids your competitors will pay per click for your targeted keywords.Even if you do all the work to estimate your paid placement costs, you will never be accurate, for several reasons:
No budgeting techniques for paid placement bids are terribly accurate, but most marketers would prefer to have at least a guess at their costs rather than shooting completely in the dark. And if you think estimating your paid placement budget is hard, just try to predict a shopping search budget. Most shopping search engines are targeted at small retailers who start by plopping down $50 on a credit card to "see how this works." Or large manufacturers that can get some cost guesses by working directly with a sales rep from the engine. No tools exist to predict the number of searches for any category of productthere is no analogue to the Yahoo! Keyword Selector tool, for example. So, although you have much better per-click cost certainty in shopping search as opposed to the bidding style of paid placement, you have no clue how many clicks you will get. So, the best advice might be to conduct a controlled experiment. As we write this, Froogle is still a beta offering (read: free). You can load your product catalog into Froogle and see how well your products seem to rank and how many clicks you get. Then you might want to branch out to a paid shopping search enginemaybe devote 5 percent to 10 percent of your paid search budget to shopping search, just to see what happens. You will not really be able to plan your budget before you start, but you can take some baby steps before committing yourself with a large investment. Set Up Your AccountsSetting up your program can often be one of the simpler steps, but it is very important. As with the other steps, you can always go back and change your setup later, but getting it right upfront can save you some time. As you have read through this chapter, you have probably made some decisions on which paid search programs make the most sense for your business. If you sell the kinds of products offered in shopping search engines, you might want to sign up and give it a try. If you know from your organic search experience what keywords your visitors are likely to search for, you have all the information you need to test paid placement for your business. You can use the tables presented earlier in this chapter to analyze each vendor that offers paid search and decide which ones seem to be the best candidates to start with, based on cost, market share, or other factors. It might be tempting to sign up with every vendor out there. Don't. The work that goes into managing a paid search campaign is larger than you think. To be successful, you must focus on your content, your per-click charges, your clickthrough rates, your conversions, and your profitability. You need to do that for every campaign with every vendor, because each one is different. Don't become overwhelmed with work before you even know how it's done. Start small. If you decide to start with just one vendor, you will probably want to use a paid placement vendor rather than a shopping search engine, because setup is usually simpler. Google and Yahoo! are the two biggest paid placement engines, so pick one of them. Choosing a single vendor also simplifies your decision on bid management and reporting toolsyou can avoid investigating and paying for a third-party tool until your program gets bigger later. (We walk you through all of your bid management choices shortly.) But how do you choose between Yahoo! and Google? Yahoo! is far simpler to budget for, as you saw earlier. Although all paid search budgets are estimates, it is much easier to estimate for Yahoo! than for Google because you can use keyword demand to estimate the number of impressions. Google's ranking algorithm relies on both your bid and your clickthrough rate, so it is hard to know how often your ad will be shown. Although it is harder to budget for Google, its ranking algorithm gives you less to manage after the campaign is up and running. Google's algorithm, which ranks listings by bid and clickthrough rate, produces less volatility than algorithms that employ bid alone, so your Google rankings will not jump up and down as erratically as with Yahoo! (meaning you can watch them less). In addition, several bidding tactics you must pay attention to when using Yahoo!gap surfing, bid jamming, and friendly URLare not used with Google, making your bid management efforts far simpler.
Snap Electronics wanted to start with one paid placement vendor and one shopping search vendor for its digital cameras campaign. They had done all the work to estimate their budget for Yahoo!, so they chose Yahoo! as their paid placement vendor, knowing they could add Google later. Even though it was free, they decided not to make the effort to sign up with the Internet Yellow Pages services, because each one would drain some of their time and they thought that relatively few people would find their company that way. They ran into their first surprise when they went to sign up for shopping search, as shown in Figure 14-4. MySimon was no longer accepting new merchants selling digital cameras. Figure 14-4. Shopping search overload. At times, so many merchants want to sell a product that shopping search engines stop accepting newcomers.Having struck out with a smaller player, Snap decided to go with one of the top-tier players, choosing Shopzilla. If you are working with an external search marketing vendor, perhaps they can set up your shopping search account for you, but Snap wanted to do it themselves without paying the vendor's hourly rate. When Snap went through the sign-up process at Shopzilla, they were pleasantly surprised to see that they could merely enter their product list into a Web form, as shown in Figure 14-5, because Snap had fewer than 30 models of digital cameras. Although they would have to figure out trusted feeds or other update mechanisms later, this certainly made it easy to get started. Figure 14-5. Signing up for shopping search. Some shopping engines, such as Shopzilla, allow product data to be entered as paid placement ads are.Setting up paid placement accounts can sometimes be a bit more complex than shopping search account setup. For example, the way you set up an Yahoo! account varies based on your situation:
If you are setting up your Yahoo! account yourself, rather then having a search marketing vendor do it for you, you will be presented with the page shown in Figure 14-6. Figure 14-6. Selecting a Yahoo! program. Yahoo! has various options for you to get started with its paid placement advertising.Reproduced with permission of Yahoo! Inc. © 2005 by Yahoo! Inc. YAHOO! and the YAHOO! logo are trademarks of Yahoo! Inc. Signing up for a Google account is not much different from Yahoo!'s process. Google accounts are self-service (similar to Yahoo!'s free program), but they cost $5 to sign up. For any paid placement account, you provide a credit card number and decide how much money to place into your account (as little as $50) and you are on your way. When a searcher clicks your ad for a keyword, the amount you bid for that click is deducted from your account. You can have the account automatically replenish from your credit card, you can have a set amount taken from your credit card each month, or you can be notified when your account dips below a certain point (so that you can manually add more funds).
Figure 14-7. Selecting Yahoo! countries. Yahoo! offers U.S. targeting by default, but allows selection of other countries during the account setup process.Reproduced with permission of Yahoo! Inc. (c) 2005 by Yahoo! Inc. YAHOO! and the YAHOO! logo are trademarks of Yahoo! Inc. You can further select location down to city and zip code if you want to refine your targeting using the geographic targeting capabilities discussed earlier in the chapter. We discuss this technique more later. Select Your Bid Management ToolNow that you have set up your paid placement and shopping search accounts, it is time to choose the software you will use to manage your campaigns, which is called a bid management tool. Bid management tools automatically adjust your paid placement bids and aggregate click and conversion data. A good bid management tool can spell the difference between a successful campaign and a failure. A few bid management tools are distributed as software you install on your own computer, but most are offered as services from Web sites. Which specific tool you choose depends on what search engines you have selected:
If you are working with a search marketing vendor, they might handle the tracking and reporting for you, so you might not need to choose a tool yourself. Twenty-five percent of search marketers choose that route. Some search marketers prefer to be more hands-on, however, even when delegating the paid search campaigns to a vendor. If you want to use the tool to see how your campaigns are doing, you might want to provide guidance to your search marketing vendor about the features you want. Even if you did not decide to work with an external search marketing vendor for your overall campaign, you might find that you would like to work with one to handle your paid search program to reduce the work for you. Here are a few that are especially noteworthy, although there are many other good choices:
For those on smaller budgets, or those who would prefer to manage their campaigns themselves, bid management tools that do not require a full-service relationship with a search marketing vendor are available. However, bid management tools are no one-size-fits-all proposition. What you need depends on your circumstances. Here is a checklist of questions you should answer before researching the right bid management tool for your program:
To make decisions on the precise features you are looking for in your bid management tool, you need to understand some of the most-used tactics in paid search marketing, such as gap surfing, bid jamming, friendly URL, and dayparting. If you plan to use any of these tactics, select a tool that supports that tactic.
Table 14-4 shows some standalone tools that do not require a full-service commitment from a consulting vendor. (Some of these tools are offered by vendors that will operate them in full-service mode if desired.) We have included the Google and Yahoo! tools in the table to help you compare.
Each tool listed in the table supports a different number of paid placement search engines and provides varying features for its automated bidding functions. Gap surfing, bid jamming, dayparting, and automated bid rules are some of the most-requested functions for these tools, so we highlighted which tools offer them. Atlas One Point and Keyword Max offer the most automation of any of the tools, with objective bidding optimization based on some metrics we explain later in the chapter, such as cost per action (CPA), profit margin (PM), return on advertising spending (ROAS), and return on investment (ROI). The other tools offer only basic rules that do not use objective analysis. Due to agreements with Yahoo!, several tools are offered as Yahoo! versions and separately priced versions for managing all of the other PPC programs. This adds to the tool's cost, obviously, but worse yet is its impact on ease of use. You cannot integrate all of your bid adjustments and tracking in one place, causing extra work for your team. The pricing shown in the table is the lowest offered for any version of the tool, but might run considerably higher for large paid search programs. Some packages limit the number of users, some the number of keywords, others the number of bid adjustments per day. Regardless, the lowest-priced version of each tool usually has some limitations. Each of these tools can raise your efficiency while reducing your workload, but they cannot eliminate human monitoring. For example, they might lower your bid to the minimum amount if it gets few conversions, but it will not drop the keyword completely, which might be the best thing to do with a real loser. Regardless, the higher your paid placement budget, and the more competitive your keywords, the more you will need a strong bid management tool. Well, there you have it. You have learned all the tasks necessary to set up your paid search program. Now it's on to step two, where we crank up our first paid search campaign by deciding our target paid placement keywords and shopping search products. Choose Your TargetsAfter setting up your paid search program, your next step is to choose the targets for your first campaign. For paid placement, this means selecting your keywords, whereas for shopping search it entails choosing the products you want to offer. The shopping search decision is simpler, so we discuss that before moving on to the complexity of paid placement. To start with, shopping search engines are limited to a fairly small set of products, so you can only offer the subset of your product line that corresponds to the product categories handled by each shopping search engine. If your site does sell some of the products offered by shopping search engines, you need to determine whether the per-click fee charged will allow your sales to be profitable. In general, shopping search engines have focused on products whose price tags are high enough that the per-click fees are worth the cost, even with a rather low conversion rate. However, you should have some confidence that your site can convert the referrals sent from shopping search engines before you sign up. If you are struggling to convert visitors now, getting more visitors who you have to pay for will not helpyou need to correct your conversion travails first. The last step is one we covered in Chapter 10getting the data to the shopping search engine. Some shopping search engines crawl your site for free, but you must develop a trusted feed for most of them. That can cost some money that eats into your profitability, but it is typically a one-time hit. Sending the data each day has a rather low operational cost after you have automated the process. Most merchants should start by sending some of their best-selling products to shopping search engines to see whether they can compete on price, shipping speed, and reliability. Will customers click through to your site above the rest? Will customers convert after they click through? Start testing some of your strongest products to see how you do, and then go from there. Later in this chapter, we provide tips on attracting the shopping searcher's click. Paid placement is a much more complicated game. Because it is based on search queries, it is not restricted to any subset of your products or servicesyou can sell anything that searchers can type. Paid placement keyword planning requires four steps:
Let's start by discussing a few tips specific to choosing paid placement keywords. Select Good Paid Placement KeywordsYou might recall from Chapter 11 that natural limits exist when planning for organic keywordsyou cannot put landing pages in place for an unlimited number of organic keywords without ruining your site. With paid placement, you can target as many words as you want. But should you? Probably not. Just as organic search has a natural limit, paid placement does, too. You will eventually get to the point that more keywords do not bring you more conversionsthey are not worth their costs. They might not be worth the per-click costs, or perhaps it just costs too many people to manage them. So although most businesses find that they can target more paid placement keywords than they target for organic search, they need to prioritize even those paid ones. If you are just starting out, it makes sense to concentrate on the keywords that might bring the highest returns. Definitely start out with transactional queries, especially those with your own brand names in them. Add informational queries that you believe will convert. (If you have done a good job tracking your conversions for organic keywords, you have an excellent head start.) One of the biggest mistakes for paid search rookies is to chase keyword demand. Just because a keyword gets high demand does not mean it will convert for you. Remember the lesson of "too hot" keywords. As challenging as those overheated words are for organic, with paid search you are paying for every one of those nonconverting clicks. Paid placement also has some advantages for search marketers who want to increase brand awareness. You can buy keywords that hit your target market for brand awareness and design relevant landing pages. For example, if you have a famous spokesperson, buy her name. Now, we know that searchers are not actually looking for your company, but if you put up a "fan-zine" site about her that happens to have a bunch of ads about your product, you might be pleased with the brand awareness you can raise at a very low cost. For example, Adidas could buy the names of the tennis players who wear their shoes. If you sponsor a golf tournament, buy golf words before the start of the tournamentyou not only get brand awareness, you also get viewers for the event. The Web site for Castrol motor oil (www.castrol.com) has a page about its spokesman, soccer star David Beckham, as shown in Figure 14-8, so Castrol might consider adding additional content and purchasing his name as a keyword. Figure 14-8. Buying brand awareness. Castrol's Web site contains information on a celebrity spokesperson whose name draws many search clicks.Remember that your search for keywords will pay off handsomely. Your friendly search engine rep will happily give you a starter list of words to buy for your industry, but don't you think your competitors get the same list? Those words are bound to be somewhat overpriced because they are the only ones being bought by any of your competitors that are too lazy to do the hard research work. This is where your iteration philosophy comes in. Continually expand your keywords to find new ones that convert at an acceptable rate. Find that keyword your competitors have not discovered yet.
Organize Your KeywordsEven if you are a habitually sloppy type, you need to think about how to organize your paid placement keywords. How you group your keywords will simplify the management of your paid search program and enable you to measure performance you could never track any other way. Both Yahoo! and Google provide ways of organizing your keywords. It would be no fun at all if they did it the same way, so we explain their approachesthey are similar. Both engines have the concept of an account, as we explained when you set up your account earlier in this chapter. Even small accounts are too large to be manageable without logical subdivision, however, so both Google and Yahoo! give you ways to slice and dice your keyword lists. Yahoo! enables you to divide your keywords into categories, which you can use to store keywords that are somehow related to each other. It is completely up to you which keywords you put together in categories, but several ideas predominate:
You can see that there are many reasons to separate keywords into categories, but luckily many of them overlap. You can decide to have categories for each individual product within a country that share the same message, for example, and you are happy to have the reporting divided on that basis, too. Google gives you the equivalent of categories, too, but refers to them as ad groups. Ad groups contain the individual keywords and the ads associated with them, just as Yahoo!'s categories do. But Google goes Yahoo! one better, by enabling you to collect multiple ad groups into campaigns. Google campaigns give you a higher-level organization that can make reporting much simpler. If you placed each related product within a country into an ad group, you could decide to do your reporting by country (by creating a Google campaign with all of the ad groups for that country) or by product (using a campaign that collected all of that product's ad groups). Google campaigns enable you to have very specific ad groups to manage your keywords within broader campaigns used for reporting. Google recommends placing keywords into ad groups that share the same messages, or two to four variations of the same message. You will want to place synonyms, for example, in the same ad group because they might share the same ads and landing pages. However you decide to divide your keywords, take advantage of what Google and Yahoo! offer you; it will simplify your reporting and your management. Next we look at another choice you make for your keywords: Which of the searchers' queries should they match? Decide a Match Type for Each KeywordAs we continue to look at how to tell the paid placement engines which keywords you want in your program, you need to specify how the keywords you choose ought to match the words that the searchers use. At first this might seem odd to you. After all, search engines decide which organic pages match searcher queries all by themselves, so why do they need your help for paid search? Paid placement differs markedly from organic search in that the search engines do not look at the landing pages to decide which ones are the best matches. They look at the bids for the keywords (and in Google's case, the searcher click rate). That's why they need your help. Besides, you want to help them, because you will increase your conversions and lower your costs. The way search engines get your help is through match types. The most inclusive match typethe one that matches the most searcher queriesuses different names on Google (broad match), and Yahoo! (advanced match), but works almost the same way. If Snap Electronics bids on the keyword digital camera using broad/advanced match, any searcher query that contains all of those words can trigger Snap's ad. Word variants also match, such as any phrase that includes the words digital cameras. So, the queries "reviews for digital cameras" and "camera for digital photography" would both match Snap's keyword purchase for digital camera. The broad and advanced match types provide the most matches for your keyword, increasing your impressions and your referrals (as well as your fees). Google and Yahoo! differ on a couple of points. Broad match is the default for Googlethe one you get if you do not select anything when you purchase a keywordbut advanced match is not Yahoo!'s default. In addition, Google's broad match also matches synonyms, whereas Yahoo!'s advanced match does not. More restrictive than broad match is Google's phrase match, which matches any query that contains the keyword in order with no word variants. So, using phrase match, the queries "digital camera accessories" and "digital camera sale" both match digital camera, but "new digital cameras" and "digital slr camera" do not. Yahoo! does not offer anything equivalent to Google's phrase match. One more restrictive match type remains, again with different names from Yahoo! (standard match) and Google (exact match). The most restrictive match type, Google exact match finds only the exact words as they are typed, so if Snap purchases digital camera with standard/exact match, no other queries will match. Yahoo!'s standard match is slightly less restrictive, also matching word variants and misspellings. Standard match is Yahoo!'s default, and it always shows standard matches before advanced matches in its ranking. Using this match type drives fewer referrals, but might give you higher conversion rates at lower cost. The final match type is called negative match by both Google and Yahoo!, and it is very important. Negative match enables you to specify words that, if found in a searcher's query, would prevent your ad from being shown. If Snap Electronics wanted to use a broad/advanced match type for digital camera but found they were getting too many clicks that do not convert for digital video camera, they could add a negative match for the keyword video to eliminate those queries while retaining more inclusive matching for all other queries. Similarly, many occurrences of digital camera might be the phrase digital camera accessories. Do you want to negate accessories because you want a lower bid for that phrase and a higher bid for digital camera? Do your homework on what searchers enter and use negative keywords to target your keywords as tightly as possible. Rather than getting bogged down in all the names, we refer to expanded match (advanced and broad) and restrictive match (exact and standard). More importantly, when do you use them? Expanded match works best for very specific keywords, such as product numbers or unique product names ("snapshot slr" or "onetouch autofocus"), where virtually any query containing those words has a reasonable chance of converting. More common words can be good candidates for expanded match if their meanings always match your scope ("snap electronics") or if their meanings are unambiguous ("honda" or "ipod" or "cancun vacation"). Longer phrases can also be good candidates if you want to capture all variationssnapshot digital camera matches "buy a snapshot digital camera" and "snapshot accessories for digital cameras." But expanded match does not work in all situations. Restrictive match is far better when your keywords have ambiguous meaningsthey can mean more than one thing. Single words and acronyms are especially susceptible:
For these ambiguous terms, you must use longer keywords (Ford Taurus or MP3 player) or add negative keywords to reduce ambiguity (Ford Motor can use Ford Gerald). One of the best strategies for match types is to start most keywords with expanded match, using your metrics to gradually identify which keyword variations have the highest conversions. When you do, you can buy those keywords using restrictive match (possibly with higher bids because they are worth more) and eliminate the original expanded keywords. An alternative is to add negative match keywords to your expanded match keywords to eliminate the low-conversion variations. Regardless of what strategy you employ for match types, keep an eye on the search engines for changes to their match typesYahoo! (then called Overture) completely revamped theirs as recently as 2004. The resources listed in Chapter 16 will help you stay informed. Decide Geographic Targeting for Each KeywordAs discussed previously in this chapter, the major paid placement vendors enable you to target certain geographic locations of the searchers who should see your ad. If you do not take advantage of geographic targeting, your ad will be shown to searchers in the country or countries that you selected when you set up your account. Geographic targeting might sound a lot like the local search concept we discussed earlier, but it is actually a completely separate way of using standard paid placement advertising to restrict your ads geographicallyshowing them only to searchers within the your area. With geographic targeting, a furniture store, for example, that delivers within 25 miles of its location can purchase the keyword furniture but ask Google or Yahoo! to show that ad only on computers within the delivery zone. Local search, on the other hand, expects the searcher to enter the location as part of the query. Search engines can restrict their results to the searcher's geography through a technique called a reverse IP lookupthey compare the IP address of the searcher's computer against a table that notes that address's geographic location. The location is usually accurate, but not always. When accurate, searchers located within 25 miles of that furniture store will see the ad when they look for "furniture" but others would not. Both Google and Yahoo! enable advertisers to specify the state, city, or even the zip code of searchers who should see an ad. With Google, you can go further, even targeting by longitude and latitude coordinates, as Figure 14-9 shows. Because geographic targeting is a relatively new idea in paid search, the various paid placement engines provide different capabilities, and they are still evolving. Use the resources in Chapter 16 to keep up with the latest changes. Figure 14-9. Geographic targeting for paid placement. Search marketers can target by searcher location, even by longitude and latitude with Google.Select Your Bidding StrategyAs with so many other parts of search marketing, you need a strategy for something as complex as paid placement bidding. Some of the concepts covered here can also be useful for deciding whether shopping search per-click fees are affordable for your product. (Shopping search engines are also beginning to offer premium services that show more information in your ad in exchange for higher per-click charges, so you need to decide whether those premium fees are worth it.) But most of what we present here is exclusive to paid placement. And paid placement is all about bidding, which at first blush seems simple enough. You don't need much smarts to able to submit a bid, or even to "win" by being placed at the #1 position. However, that is not really winningwinning is driving conversions that cost less than they are worth. That kind of winning demands a bidding strategy. Rookie paid search marketers typically have no bidding strategy, or at least they do not know they have one. But the old saying "even no decision is a decision" applies here. If you do not make an explicit choice about your bidding strategy, you are basing your decisions on something, but you are just not articulating what it is. Most "strategy-free" search marketers run through a series of strategies, moving on to the next one each time they get burned by their current one. Table 14-5 shows a range of philosophiesthe ones at the top of the table tend to fixate on high rankings, traffic, and other measures of success, regardless of budget. In fact, two-thirds of all sites measure themselves based on nothing but pure traffic. The strategies at the bottom of the table tend to focus on budget constraints and efficiency rather than effectiveness. Most paid search marketers will benefit from one of the middle paths.
Newbie search marketers often choose a goal at the top or bottom of the table, without a great deal of thought. Sometimes these choices can be disastroustwo marketers both pursuing the "top dog" approach can burn through their budgets in a few days. Most of the time, poor choices are not so awfulmarketers who look for bargain keywords at low prices hurt their company's return on their marketing program, but it is not a tragedy. To be fair, those pursuing top rankings do have a reasonsyndication. Paid placement ads are syndicated to partner sites, as you saw in Chapter 2. AOL, for example, shows Google's paid placement ads on the AOL search results page. However, AOL does not show all of the adsthey show just the top three or four. So, those that are constantly jockeying for the top positions do have their reasons; but if the top spot costs more than it is worth, it is not helping them. A little thinking will lead you away from approaches at the extremesthe top or the bottom of Table 14-5. It is rarely as critical as top dogs think to be #1 in the paid placement rankings as opposed to #3. Likewise, skinflints that stay within their budgets without tracking any return on that investment cannot succeed at anything. Over time, successful marketers move to the middle, which some experts call objective biddingusing metrics to drive your bids. So what metrics are there? Lifetime value is one way of quantifying what each new customer is worth. Direct mailers have used this metric for years to decide how much marketing investment to make in a customerthey do that by deciding what a new customer is worth, not just for the first purchase but for the lifetime of purchases from that customer. If your company can calculate the lifetime value for your customers, you can take the most long-term view for your search marketing. Most companies are not so forward-looking, however. The next-best way of calculating the value of each conversion is its profit. If your company can calculate the profit on each conversion, a strategy of maximizing profit within your budget will probably be best. The problem with maximizing profit, however, is that it ignores the reality of a budget. For some companies, as long as you maximize profit, they will promise you the required cash flow to keep spending. This is good business, but it is a relatively rare approach. More typically, your paid placement program will have some kind of budget constraint, and your job will be to maximize your returns within that budget. As you might expect, the best place to start is probably to maximize your profit within your budget. Some companies track profit marginthe percentage of profit per dollar of revenue. But not every company can track profit, either. Most companies track success based on metrics built around sales revenue or advertising budgets (or both), rather than profit. Because your company's goal is to maximize profit, operating based on revenue and budgets is not necessarily wise. If your company already has a standard way by which it measures all marketing spending, however, you are best advised to use that same method for paid placement. By doing so, you allow paid search to be compared to other forms of marketing, which typically shows why search is such a good investment. Because none of these other techniques maximize profit, you always need to be careful in placing too much faith in them. Still, they are frequently much easier to calculatemany of the bid management tools we showed earlier can calculate them automatically and even adjust your bids accordingly. So before we show you these magical formulas that enable you to place your paid placement campaign on autopilot, we once again warn you not to be seduced by them. If you can figure lifetime value or the profit on each conversion, that is great. Unfortunately, many search marketers cannot calculate these measures, so they need other approaches:
Table 14-6 shows how different strategies for objective bidding can result in very different outcomes. In each case, a bidder optimizing for one of these metrics can succeed in returning a profit, but a different strategy that optimized none of these metrics returned a higher profit.
Understandably, if your company uses one of these metrics to judge the value of all marketing spending, it behooves you to do so, too. A variation on optimizing for a single metric is to use a self-funding strategy, which always ensures that your sales are profitable while possibly allowing you to reach higher sales volumes (and thus higher overall profits). To pursue a self-funding strategy, you start by calculating your allowablethe highest CPA that you will pay for a particular product. After you know your allowable, you can use the conversion rate to determine what your highest per-click bid should be, as shown in Table 14-7.
Using Table 14-7, Snap Electronics was able to pinpoint its optimal bid for its digital camera keywords. For example, Snap decided that it was willing to pay $50 for every sale of a Snap Shot digital camera$50 was Snap's allowable. (In real life, you would have a different allowable for every model of camera, so that you would allow more spending for models that returned higher profits.) Expecting to convert 2 percent of all keyword referrals into sales, Snap's opening maximum bid for digital camera keywords was set to $1. We keep harping on the point that optimizing your bids for any of these metrics does not maximize your profit, but you can at least experiment with your allowable for each product to increase your profit (if you are able to calculate your profit in the first place, of course). You can increase or decrease your allowable, letting your bid management software adjust your bidding, and you can keep measuring your overall profit until you seem to have an allowable that maximizes profitor gets it as high as you have been able to, at any rate. Economists refer to this as testing the elasticity of your marketstretching your allowable to see when profitability declines. Although not perfect, this strategy improves your overall profits without requiring expensive manual bid management. Make Your BidsFinally, let's get to what you have been waiting formaking your bids for your paid placement keywords. If you have adopted an objective bidding strategy and analyzed your numbers, you can either use your bid management software to control your bidding, or you can do it manually. For large campaigns or for competitive keywords, you need to rely on automated bid management to succeed. The minimum bid that you can place on a keyword ranges from 1¢ to 10¢, depending on the search engine, and every bid must be at least 1¢ higher than the bid below it. The preceding section showed how you can use your allowable and your conversion rate to calculate your maximum bid. Except it is not so simple. The maximum bid we calculated is just a starting point. Until you start buying keywords and you track your conversions, you do not know whether your conversion rate really is 2 percent. And even if you do average a 2 percent conversion rate across all of your keywords, some will undoubtedly convert at higher rates than others. Snap Electronics struggled with this problem at the start of its paid search campaign for digital cameras. They wanted to start with the most popular keyword, digital camera, so they could drive the most conversions to the site from the start. They knew that they could bid $1 for that keyword, and they remembered when they did the budget exercise back in Chapter 8 that the top spot went for just 85¢ per click. Unfortunately, a few months had passed while getting their program approvednow it was Christmas season. Dozens of new competitors were bidding up this overheated keyword. Figure 14-10 shows the top Yahoo! bid has escalated to $2.50 per click! Snap would have to get an impossible 5 percent conversion rate to make a profit, according to its calculations from Table 14-7. Snap reluctantly decided to move on to less-popular keywords. Figure 14-10. Making your bid. Yahoo!, like all paid placement engines, provides a simple way for you to make your maximum bid for any keyword.Reproduced with permission of Yahoo! Inc. © 2005 by Yahoo! Inc. YAHOO! and the YAHOO! logo are trademarks of Yahoo! Inc. Because Snap cameras get great reviews, Snap decided on the keyword digital camera reviews. Snap was excited to see the keyword had just a 35¢ bid for the #1 spot, although they knew it would drive far less traffic than digital camera. Snap quickly entered a maximum bid of $1 and took the top spot, knowing that they would never pay more than a penny higher than the #2 bidder. Unfortunately, a few days' worth of data revealed that the conversion rate for that informational keyword was distressingly low, less than 0.5 percentit was not worth what they were paying. Snap was forced to temporarily withdraw the bid and to focus on other keywords that would convert at higher rates. Fortunately, they found that their branded keywords, such as snap digital camera and snapshot camera converted at more than 2 percent and their competition for those keywords was decidedly lessSnap took the top spots for less than 25¢ each, providing very strong returns. Snap was smart to start their paid placement campaign by testing with a very small part of their budget. As soon as they got some results, they iterated their bids to go after new keywords that paid off. They will continue to do so as they gain experience. Besides setting your maximum bids, you can also set budget maximums (either on a weekly or a monthly basis), so that you do not spend more than you should. If Snap Electronics had followed through on buying a popular keyword, such as digital camera, it would have drained its monthly budget in just a few days. Instead, Snap could set a daily maximum spending rate to spread out its budget throughout the month. The search engines accomplish this magic by rotating Snap's ad among other ads, so that even if Snap had the #1 bid, their ad would not be shown to each searcher with a matching query. The ad would be shown just enough to keep spending under the "cap" for the day. Sounds great, right? It can be, when used properly, but spreading out your budget by letting the search engine randomly show your ad is no way to maximize your conversions. You would be better off taking other actions to stay within your budget:
Daily and monthly caps do have one good usethey protect you from sudden shifts that could be disastrous to your budget. Suppose that you were running the search marketing for Hilton Hotels in France. What do you think happened to your paid search budget the day that celebrity Paris Hilton burst upon the scene with a naughty video? If you had the top bid for the keyword paris hilton, you might have gotten a lot of erroneous clicks to drain your budget before you figured out what was going on. A daily cap could stop the bleeding. Caps can also save you from yourself. Paid placement requires precision in keyword selection, bidding, and ad copy. Do it long enough and you are bound to make a mistake that drains your budget without any conversions. Caps can stop a run on the bank. In Google, you can set caps by campaignYahoo! has caps only for your whole accountso you can put lower performers in their own Google campaign and turn them on and off depending on how the rest of your campaigns are spending your budget. But don't use caps to stay within budget when you could be making your spending more efficient. It's the lazy way to keep from blowing the budget, but it hurts your return. No random ad placement by search engines can drive as many conversions as your active management based on metrics will. Use caps as a firewall against disasters, not as an operational strategy. Attract Searchers' ClicksTake a deep breath! You're getting there, but you still have a little more to learn. It is not enough to set up your program and bid on the right keywords. Just as a direct-mail piece needs the right words to get you to open the envelope, you need to get searchers to click your listings. Effective copywriting is crucial for driving clickthrough rates for your paid placement listing, but for shopping search it is a different game. We look at both, starting with paid placement copywriting. Optimize Your Paid Placement CopyNineteenth-century American humorist Henry Wheeler Shaw once said, "Money will buy you a pretty good dog, but it won't buy the wag of his tail." And so it is with paid placement, where your per-click fees can buy impressions for your ad, but only your copy will get the clicks you need from searchers. You recall from Chapter 11 how difficult it can sometimes be to craft pages that contain the search terms (so that both the search engines and the searchers find what they are looking for). With paid placement, you will have no such difficultyyou can lead searchers to whatever pages you want with your paid ads. Well, almost. Paid placement engines do have editorial guidelines designed to protect their business. Each search engine has somewhat different rules, but they are very similar:
Both Google and Yahoo! publish extensive tips about how to follow their guidelines and how to make your ads as successful as possible. There are three critical parts to paid placement ads, as illustrated in Figure 14-11:
Figure 14-11. Anatomy of a paid placement ad. All paid placement ads have the same three components, as this diagram from Google demonstrates.Source: The Maximum Effect p. 26 Google's ads are shorter than Yahoo!'s, which makes them a bit more challenging to write, but some of Yahoo!'s syndication partners truncate the ads, so you ought to put your best stuff up front. The key advice for creating good titles and descriptions are the same for any paid placement ad:
After you have decided on your wording, it's time to actually place your ad. It can save time to use the same copy for an entire Yahoo! category or Google ad groupthat is why you collected them together in the first place. You can then edit each keyword individually to fine-tune the copy, as shown in Figure 14-12. Figure 14-12. Editing titles and descriptions. Yahoo! shows exactly how your ad will look when presented in Yahoo! and the other syndication sites.Reproduced with permission of Yahoo! Inc.c. © 2005 by Yahoo! Inc. YAHOO! and the YAHOO! logo are trademarks of Yahoo! Inc.
We have focused on titles and descriptions, but the last part of a paid placement ad, the display URL, is also important. Display URLs tell the searcher where their click leads them. If you have a well-known brand or Web site, make sure that the display URL plays it up. Manage Your Paid Placement AdsWe have continually counseled you to go after as many keywords as you can profitably afford, and we have told you to place keywords in every ad, and to test variations against each other. How can you reduce the labor for all this work? Let's start with inserting keywords into hundreds or thousands of ads. You have probably set up Google Ad Groups or Yahoo! categories for keywords that are similar in meaning. That's great, because those keywords can share the same ads, reducing your copywriting work. But it causes a new problem, because you want the keyword the searcher entered to appear in each ad. If you have used the same copy for each ad, how do the keywords get inserted? With Google, you can use a clever tool called Dynamic Keyword Insertion (DKI). When you define your titles or descriptions, you can use curly braces to indicate the place where you want the keyword inserted into your copy. Snap Electronics has two dozen models that they placed in a single ad group so they could share the same copy for their ad across all models. But they wanted the model name in the title, so they coded their title as "Buy { Keyword: SnapShot Cameras} direct"which inserts the searcher's keyword dynamically. For example, if the searcher entered "snapshot slr," the title of the ad would read "Buy snapshot slr direct." By controlling the match type of the ad group, Snap ensured that only phrases with SnapShot were part of this ad group. Sometimes, however, the keyword entered by the searchers are too long to fit in the title, such as "snap snapshot slr x900." In that case, the default text after the "Keyword:" in the braces would be inserted instead, creating a title of "Buy SnapShot Cameras direct." Google's DKI can insert keywords in the title or the description, enabling you to raise your click rate without much extra work. Yahoo! and other paid placement engines do not offer any comparable tool, but they usually accept ad content submitted using spreadsheets, so you can use macros in your favorite spreadsheet to insert your own keywords into standard ad copy. You will have to code your macro to count the number of characters to ensure you do not exceed the length restrictions for the engine you are feeding (substituting default text in that case). It is not as elegant as Google's capability, but can be equally effective. Whatever it takes, make sure that you get the searcher's keywords into every ad possible. It dramatically raises your clickthrough rate. However, it is not the only way to raise your click rate. You can also test multiple versions of ad copy for the same keyword to see which one draws more conversions. As you might expect, this can be labor-intensive, too. So-called "A/B testing" (because you test version A against version B) can be automated by many bid management tools. Google offers a feature called auto-optimization that enables you to rotate up to six different versions of ad copy, with Google eventually settling on the one with the highest click rate automatically. Some advertisers prefer to test their ad copy themselves, because they want to conduct more extensive sampling than Google does or because they want to select the winning version based on conversion rate rather than clickthrough rate. Regardless of how you manage your ad copy, finding ways to raise your clickthrough rates with the least possible effort is a winning strategy. You can also work on your shopping search listings to attract more clicks, as you will see next. Fine-Tune Your Shopping Search DataOptimizing your data in a shopping search engine is the key to attracting clicks from searchers. Figure 14-13 shows why shoppers use shopping search enginesyou need to satisfy those needs to persuade them to click your listing. Figure 14-13. Shopping search uses. Shoppers like shopping search engines for many different reasons.Source: ForeSee Results (December 2002) To attract clicks on shopping search engines, you must concentrate on three major factors:
So, pay attention to your data to ensure that your listing is found, and pay attention to the ranking factors to be placed at the top of the list. Ranking at the top is importantone leading shopping search engine reports that 50 percent of all shopper clicks go to the top five listed merchants. When first starting out in shopping search, you might need to discount your price to attract shoppers until you boost your merchant rating. Some shopping search engines allow you to pay higher per-click fees to become a "featured" merchantNexTag always ranks by per-click bids, just like a paid placement engineso that is another way to be listed first in the search results.
Optimize Paid Search Landing PagesWe covered organic search landing pages in detail in Chapter 12, and many of those tips apply to paid search landing pages, too, except that organic pages must appeal to search engines to be found in the first place. Paid search landing pages need appeal only to the searcheryou can optimize paid search landing pages to drive your visitor to a conversion. For paid placement, a few editorial guidelines are insisted upon by paid placement engines:
These guidelines (and the rules we discussed earlier for the ads themselves) are not very hard to follow, but the search engines are quite serious about them. Catherine Seda, in her book Search Engine Advertising, reports that Yahoo! editors reject as many as 30 percent of all ads submitted to them, many for relevancy. Because Google's ranking algorithm rewards higher clickthrough, it does not need to spend as much time policing ads and landing pages for relevancy, although it rejects many ads, too. As with organic search landing pages, you must make it simple for visitors to convert. If you reinforce what searchers are looking for (their search keywords) and why they clicked your pages (your ad copy), you will likely get a good conversion rate. Remember that you might have several pages between your landing page and your conversion page, so make sure that the entire path is designed with conversion in mind. For shopping search, the path is usually much shorteryour shopping search landing page is typically the "buy" page for that product on your site. Whether using paid placement or shopping search, your landing pages (and your entire conversion path) requires careful design and, above all, testing. Just as you test everything else about paid search, test your landing pages, too. Test the amount of copy on the page, the size of the product images, the kind of language used, the links to background information, and more. To test, you need to measure your conversions, which we discuss in Chapter 15.
Measure and Adjust Your CampaignsHey, you have a campaign up and running and your bid management tool is set up, so now you can just wait for the money to roll in, right? Well, not exactly. If you recall, your strategy is to iterate, so it's time to check out how you are doing and then make changes to improve. Bid management tools are very helpful, but they cannot substitute for your good judgment. You must monitor new campaigns quite closely, because you are learning a lot. Over time, each campaign requires less monitoring, although nothing stays static forever. In Chapter 15, we talk about how to keep campaigns on track after the initial adjustment period, but now let's focus on how to track your campaigns as they come out of the gate. Start out by tracking your campaign according to your objectives. If you are tracking profit, see whether it is meeting your projections. If all you can measure is your clickthrough rate, well, at least check that. And watch your bid rates for paid placement. If competitive bids are rising and your position is falling, it's time for deeper analysis. Measure as granularly as you can. Track your paid placement results by keyword within search engine within time of day, if possible. Examine your shopping search conversions by product within search engine. Notice which targets are performing and which are not, according to the metrics you have chosen. When your paid placement keyword ads or your shopping search listings are not being clicked, take action:
Note that Google will disable any ad not delivering at least a 0.5 percent clickthrough, so it pays to scrutinize results regularly to optimize. (In the case of broad- and phrase-matched keyword variations, the clickthrough threshold might be higher.) Other paid placement engines might not be so stringent, so you might want to experiment with other paid placement engines if your ads are constantly disabled by Google. What if you are getting clicks, but not enough conversions? Again, look at when those paid placement keywords or shopping search products do perform well. If you catch them performing under certain conditions, dayparting and other restrictions on your keywords might be appropriate. Limiting your product catalog or restricting it to certain shopping search engines might make sense. As above, you might also want to experiment with new keywordsespecially more specific keywords that might better target your audience. If those measures do not work, try some other ideas:
Oh, and one more thing. You should do everything listed above when your targets are performing, too. That is how you follow your strategy of being more efficient. It is far more likely that you will double your conversions by improving your already strong targets than by buttressing your weak ones. And getting more impact out of your winners is not any harder than fixing your losersin fact, it is often easier. If you make your winners truly efficient, you will do well to spend more money on them and drop the targets that are borderline performers. Snap Electronics had a lot to learn when they began purchasing paid placement keywords. You recall that back in Chapter 7, "Measure Your Search Marketing Success," Snap wanted to purchase the keyword digital camera, but earlier in this chapter they found themselves priced out of the market. Reluctantly, they decided to experiment with the keyword digital camera reviews instead. Based on their allowable of $50, Snap needed to average a conversion rate of about 2 percent to justify a bid of around $1, but we saw that the top bid was only 35¢. Snap started by trying to top the current #1 bidder, knowing they would need to approach a 1 percent conversion rate if the price went up. Well, the price did go up, because the previous high bidder topped 70¢, jockeying back and forth with Snap as the average per-click fees for Snap's campaign hovered in the low 70s. But within a few days, Snap saw that its clickthrough rate was way too lowless than 1 percentso the conversion rate was minuscule, far less than the required 1.5 percent. Snap was forced to temporarily abandon this keyword to diagnose the problem. Clearly the first problem was the low click rate. Snap took a close look at its ad copy (shown as "Ad A" in Table 14-8) and decided changes were needed. Snap quickly revised the ad copy to "Ad B" to try to boost its click rate. The results were dramatic, as clickthrough jumped to 5 percent, but the conversion rate was still less than 1 percent, much less 1.5 percent. Clearly, unqualified searchers were now clicking and then abandoning, so Snap set out to change the copy again, to what is shown as "Ad C," in an attempt to attract only those with a propensity to buy. It worked. Although click rate went down, conversion rate went up, but not quite to the magical 1.5 percent they needed.
What now? At this point, they examined their metrics, and found that many of their clicks were for queries such as "cheap digital camera reviews," "bargain digital camera reviews," and "discount digital camera reviews." They also found that some of the keywords contained the names of competitors ("nikon digital camera reviews) that drew few clicks and next-to-no conversions. Digital video cameras was a keyword that brought many clicks from hurried searchers, but near-zero conversions. By ferreting out a list of additional words that correlated with a lower conversion rate, Snap was able to add negative terms and increased their conversion rate to more than 1.7 percent. Not satisfied yet, they began to look at dayparting to see whether their conversion rates have any pattern to them. And the beat goes on. |