Brand keyword conversions are not a given.

I have heard all too often things akin to “Sales from branded keywords are not incremental”, or “why should an SEM be rewarded for sales coming from the Brand? The Brand did all the work.” This portrays a very myopic view of SEM. In fact, it might as well be called “Search” and leave off the “Marketing.” As much as we want to believe ourselves separate and apart from the rest (perhaps some think, 'above'?), we are not. We are (or should be) first and foremost “Marketers.”

On the very surface ask yourself, “where in the purchase funnel is the searcher when they enter a branded keyword?” If they enter “ditech mortgage”, does this necessarily mean they have already decided to buy ditech? No. Compound that with the fact that “diteh home loan” will likely provide a very different type of user. Unless the ‘brand’ converts 100% of the prospects who use a branded search, the 'brand' is not simply 'selling itself,' there is more work needed. So, how do you improve conversions?

A good SEM works very closely with the client to understand the relationship between the search term, ad copy, destination URL, competitive set, closing offers, buy flow, LTV and much more. Fore instance, Atlas released a study showing that conversions from clicks to sales are proportionately lower during certain parts of the day (this concurred with our own studies). So, what do you do about it? Should you provide a different experience between 8am and 11am vs 11am and 2pm? How should you manage your bids? If you limit your perspective to “brands sell themselves” you will likely never ask questions like these, never mind dedicate the resources needed to answer them.

Last year we met with a partner of ours for whom we are doing a specific acquisition program. During the course of the conversation, they mentioned some of the things they were doing on there core search program with the (then) current SEM. We made some minor suggestions and observations, which they then implemented (they did not mention this until much later). After doing this, their results improved significantly. They are now working with us on their corporate program as well as our acquisition program.

Now, the reason we were able to propose the things we did is because we did not assume the brand simply sold itself. In our own program for the partner, we have (and continue to) tested through multiple scenarios, through the complete buying cycle (we do not pass off the user, we have a closed loop process) and were able to provide ‘off-the-cuff’ direction that was very beneficial to them. Only by questioning the assumptions do we find the answers that lead to more conversions.

Well, at this point one may ask, shouldn’t all good SEMs do this? I will say that most want to. But, if your compensation is fixed to spend, and the client has only a set spend budget, what kind of resources can you really afford to dedicate to total program analytics? Can you afford dedicated experience experts, research experts, or fully staffed analytics groups? No matter how well the brand is doing, assume it can do better and behave accordingly. Your clients will see growth they did not know was available and you may reap the rewards, if you structure your compensation properly.

With Google’s Preferred CPC sit back, Relax, and… maybe not

What if I told you that I could guarantee the price you'll pay for every dinner you eat over the next month? This way, you can set your dining budget and not have to think about the cost of each meal.  What I won't tell you is what the meal will be, if you'll get an appetizer or dessert nor how long you'll have to wait in line for a table. My guess is you would say "No thank you."

Google's new Preferred CPC is not that bad, but it can be if you're not careful. That means it is far from the "set it and forget it" solution. Since the value of a click is influenced by more than the CPC, and the influences can change while the CPC remains relatively constant, you can hardly afford to ignore the campaign for any length of time.    

Ultimately, you want to manage to the cost / profit per specific event (Conversion usually) and you'll want as many of those events as you can get.  If competitors increase their bids or improve their quality score, you will loose position and your CTR will drop, giving you fewer leads. If their offer improves you will likely see lower conversions. So, while your CPC remains constant, you're volume will drop and you pay more for each conversion you do keep. Kind of makes you wish you knew what was on the menu, huh?

If you have limited resource and the Preferred CPC tool is very appealing to you, consider two tools, which when combined will serve as a daily warning of any metric-changing events.  

First, created auto reports in Google and have them emailed to you each morning. They should include past 30 days and yesterday’s numbers for: Impressions, Clicks, CTR, Position, Cost and CPC (though this should be stable). You want to look for any sudden changes, or trends that will take you outside you're target metrics.
  
Second, sign up for adgooroo.com and set auto alerts to be emailed to you showing new ads and new competitors (preferred bid or not, this is a good tool.) This will alert you to competitive pressure that may hurt conversions.

Finally, do or quick calculation and divide the orders for yesterday by the clicks. If the conversion is different than you expected, dig deeper...something changed. 

By now, you’re probably asking, “what’s the point of the preferred CPC if I have to do all this?”  Well, that is the point. The CPC is not the key metric here. The amount you’re willing to pay depends on many factors, all of which you’ll need to monitor no matter what. This is not to say that the Preferred CPC does not have a place. If all your other metrics are relatively stable, or you are driving traffic with only the cost per click as a prime metric, then this will work for you. If, however, you manage your program to a profit metric, using the Preferred CPC is moderately beneficial and must not lull you into a comfort zone. SEM is just not that simple.

Yet more protection we don't need, from Utah

This week, Utah, under the guise of protecting us poor, ignorant consumers, has taken a step (which will likely be reversed) to prevent businesses from diverting our online attention by making it illegal for anyone to bid on a trademark term they don’t own. We are not talking about using the trademark in the ad, just showing an ad when the trademark is searched.

In rationalizing this law, Utah House Majority Leader David Clark likened the bidding on others’ trademark terms to diverting a shopper who goes to a particular department store to buy a dress shirt. He said, "You get right to the front door and somebody whisks you away to a different store."

His analogy is a bit flawed…okay, a lot flawed.

A more accurate analogy might be: I intend to drive to Menards (home improvement store) to get paint, and on the way I see the sign out in front of Home Depot advertising a sale on paint, so I go there.

Or, a bit further removed, Burger King sets up shop across the street from McDonald’s knowing that they will benefit from diverting some of the Mc D’s traffic.

I believe our country is what’s been called by Milton Friedman a free and open market (at least as close as we can really get). Given how “republican” Utah is, I’d think the state legislators might have something better to do than inhibit free markets and communication.

If Mr. Clark is really serious about not letting one company divert our attention during the purchase process, then Utah ought to force businesses to remove all exterior signage. Goodness forbid, we may actually find a better deal by reading advertising along the way.

Managing Contextual Ads

Rant on why there is only one real contextual ad method

Running a contextual advertising campaign through the big three can be very easy, click and go. They make it simple, fast, and if you’re not paying attention, expensive (relatively). Even with Smart Pricing from Google, you can pay too much. In these scenarios, your ad CPC is the same as or directly related to your search CPC (Smart Pricing adjusts it downward). First suggestion: Don’t run a contextual ad buy mixed in with your search program – more below.

Before you start contextual advertising, identify your metrics. For some, the desired event is exactly the same as the search program, a sale, newsletter signup or other asset building event. For others, it may be less lofty, such as engagement metrics which they believe will be a precursor to an asset building event; something like page views or time spent, or an information request. This is not just good marketing, it sets the stage for keeping search and contextual apart. Only after thinking about this for some time do people begin to create a separation between contextual advertising and search. Even though these can be run through the same campaign / ad group, they are actually two distinct vehicles and the metrics should treat them as such. Which brings me to my next point…

To properly manage a contextual ad program, you should create an independent or separate campaign (or at least a separate ad group.) The new campaign will allow you to set a max budget and target independently from the search campaign. This has some advantages when it comes to optimization.

With proper tracking, you can more easily measure your results and manage the program against the content target metrics. You will also be able to finitely control the budget that the contextual program can absorb based on how well it is performing. By keeping it separate, you create a necessarily separate focus on the relevant metrics and value.

Since the intention of a person that sees your contextual ad is likely different than that of a searcher, the ad should treat them differently. Rather than a message used for search that may say “buy now”, your content ad can be more targeted to a content audience with a “more information” message for instance. You’re hitting someone in a different stage of the purchase funnel, so cater your message accordingly – then test it.

Like search, the content of the ad, and the landing page should be as specific as possible. This will allow the content matching to more closely align your message with the information the user is viewing. Allow some ad groups to remain broad, such as “home entertainment systems”, while others get progressively more targeted down to the models (if appropriate). Populate the ad group with the relevant keywords, allowing for some but limited duplication at the more general keyword level and reserving model or product specific keywords for the ad groups with appropriately targeted copy.

Finally, the landing page should be as targeted to the ad copy as possible. Bring the user deep into your site if the subject becomes more relevant. This helps keep the user engaged by tightly connecting your content to that which triggered the ad. If someone is at the stage of researching a specific product, and your ad speaks to it, but you bring them to a general category page, you’re taking them back several steps in the purchase process. In their eyes, they’ve already covered this material. You’re job is to add value to the research process, not set it back a step. So, help them by keeping their progress going and making the landing page well targeted.

Don’t let the similarities in considerations between search and contextual advertising mislead you. These considerations, at some level, are the same in all marketing efforts. It is what you do with them that will distinguish the vertical. Only by virtue of technology is contextual advertising related to search advertising. From a consumer’s perspective and therefore from a marketing perspective, these are two entirely different beasts. From messaging to landing page to desired event, these two are distinct. While it may be easier to lump them into a single buy, it is far less effective. Keep them separate in thought and practice.

Contextual ads – from rant to recommendation

Google is search, right? Well, before they got into tv, radio and print, they were also ‘not search’. They were, well ‘seek’. In other words, rather than wait for the user to search for you, you would ‘seek out’ the user. This is popularly known as contextual advertising.

This is the area between search advertising and traditional online advertising. When it was launched (with AdSense publishers in 2003), it was based on the Google algorithm which compared ad content with publisher site content and tried to make a match based on how closely the two appeared to be related. To make up for the difference in relevancy and proximity to actual conversion (search being deemed much closer to the desired action than content), Google implemented Smart Pricing. This adjusted the CPC for contextual downward in proportion to the click value based on Google’s calculations.

The reality is there are a lot of content players today. MSN / Live and Yahoo! are have their versions. Quigo has Ad Sonar, Vibrant Media, Pulse 360 (Kanoodle) and more are out there. Unfortunately, since Google launched the contextual advertising rocket, there have been different and not so well related notions of contextual advertising. Randy Schwartz covered this a couple of years ago, outlining four types. His types make sense. However, I am not so sure that I accept that these are all indeed contextual advertising.

In my mind, contextual looks at the context, or content, of the page and then based on that decides if the ad is appropriate. This is different than picking a site about a general subject matter. For instance, if I sell baseball equipment, I’d rather run on a news article page about the home town baseball team (even if the page itself is under the subject local news), than on a sports site’s page about the Stanley Cup Playoffs. This is the difference between content / context and subject / category targeting. I make this point to lend context (no pun intended) to what follows.

Type of Contextual Advertising
Algorithmically based: The first, and in my opinion only true, contextual advertising method is that based on the actual CONTENT of the page. In this method, the algorithm looks at all that is on the page, looks at your ad content (landing page too, if done right) and decides if there is a match. In the more advanced methods, the algorithm looks for negative associations and does not deliver your ad (ie a car ad on a page about automotive recalls).

Keyword Based: Another, and in my mind poor, method is inline contextual (contextual being something of a misnomer) keyword based. It looks for keywords in the body of a page, creates a hyper link on the keyword and then either displays your ad on the mouse over with a hyperlink or sets a hyperlink though to your page. I have seen words highlighted for products that were in fact related to the word, but had absolutely nothing to do with the page content (car filters linked to room air filter, Home entertainment content linked to CA (B2B IT) based on the word ‘Clarity’).

Category Based:  Advertising is base simply by selecting content categories (like finance, automotive, sports, etc) on which your ad will appear. While the content may be relevant, this is just like everything else we’ve done online. Select your target audience, find what sites they like (or types of sites), and advertise there (okay, over simplification, but you get the point). Google set this up (properly) as site targeting, rather than contextual advertising. Others are pushing this as contextual advertising. The problem is your ads always appear regardless of the strength of the connection.

Behavioral Based: This has somewhat been taken away from content and spawned a category unto itself as it should be. I won’t elaborate – its just behavioral targeting, or BT.

The recommendations for Managing Contextual Ads are based on the algorithmically based method. The others have a place, but they are not really contextually based in the pure sense of the idea.

Recommendation =>