Search marketing has become incredibly sophisticated and complex over the last 15 to 20 years. But while the sheer number of options and settings can be overwhelming to marketers the fundamental ingredients have stayed the same. The best search marketing happens when you have the:
- Best keywords
- Best account structure
- Best ads
- Best landing pages
- Best bids
When these five ingredients are brought together the results can be outstanding. We have seen clients increase their spending from $2,000 to $200,000 per month whilst increasing their profitability through the application of these techniques. In this article we will discuss each ingredient and describe succinctly what ‘best’ looks like. We’ve had to miss out a lot the details but this overview will give you that top level picture for what you’re aiming for.
Best Keywords
Obtaining the best keywords is not easy, but the goal is simple: Full Coverage. Full coverage means having every relevant keyword, from the largest ones (by impressions) to the smallest. How small? When keywords generate too few impressions they are tagged as ‘low search volume’ by Google and Bing. This tag identifies keywords where search volume is so low that these keywords don’t enter the ad auctions. Keywords that can’t generate ads are irrelevant so this tag defines for us a useful lower bound and allows us to define our goal better as:
Full Coverage = Marketing with every relevant keyword that isn’t tagged as ‘low search volume’
Now we have a definition the next task is to find these keywords and ensure they are relevant. There are broadly two approaches that advertisers use:
- Supply approach: Advertisers review their supply in terms of products, services or landing pages and think of ways people may search for them. There are two challenges:
- To discover all the ways people search for their products without leaving big keyword gaps
- To not flood their accounts with low search volume keywords where the keyword is relevant but doesn’t buy any traffic
- Demand approach: Advertisers use keyword discovery tools to find search terms which they turn into keywords. Often these tools rely on the search suggestions we use every day on search engines as shown below. The key challenge of this approach is understanding the intent of each suggested keyword and matching it to the products and services of the advertiser.
The demand approach is by far the superior of the two, and is the only method that can give you confidence of achieving full coverage. An added bonus of the demand method is that it also finds search queries that don’t match the supply of the advertiser. These search queries can be added as negative keywords up front and save much time and expense compared to manually mining search query reports to weed out the bad ones.
Best Account Structure
Creating an optimal account structure starts with defining the meaning of ad groups: what are they about and who they target. Once this is determined the task of clustering ad groups into campaigns and accounts is of secondary importance. Any clustering that works in terms of efficiently applying campaign level settings or easing the creation of reports will do. In deciding the meaning of ad groups the goal is twofold:
- To differentiate between users of different intent so they can see more relevant ads
- To separate higher performing users from lower performing ones so that bids can be set more optimally
Intent defines the meaning of the search query. If the meaning of two search queries is the same then they have the same intent and can be acquired in the same ad group with the same set of ads. If the meaning of two search queries is different they should be acquired in two different ad groups, with different ads that speak to each specific intent.
The second reason for a more granular account structures is to separate users by their performance. If we can cluster high performing users into one ad group and lower performing users into another we can bid more optimally than if they are acquired into the same ad group with averaged bids. Determining how to cluster users by their performance is an important exercise and the right method will vary considerably from market to market. The more common performance drivers we see are based on the user’s physical location, their language, audience, or device.
Search query control
An often overlooked part of account structure is search query control. Search query control is the task of ensuring that any search query ends up triggering just one ad group – the one that is most relevant to the search query. Without search query control a single search query is eligible to trigger ads in up to 10 ad groups or even more.
An illustration may be helpful here. Suppose you have two keywords in two different ad groups. ‘+shoes’ and ‘+blue +shoes’. A user typing in ‘cheap blue shoes’ will be eligible to trigger ads in either ad group. However if we add the negative keyword ‘-blue shoes’ to the first ad group the user will only see ads from the second ad group. That ad and landing page would be specifically about blue shoes and therefore be a better match for the user. This type of negative doesn’t try to block traffic altogether, but rather it steers the traffic to the most relevant ad group. We call these types of negatives ‘cross negatives’, because the creation of one keyword say ‘+blue +shoes’ triggers a corresponding negative keyword ‘-blue shoes’ in another ad group. If for whatever reason the keyword ‘+blue +shoes’ is removed (e.g. because it is low search volume), it is important that its corresponding cross negative is also removed.
Best Ads
Having the best ads starts with having a great account structure with search query control. That’s because it is essential that the ads in a given ad group are buying traffic with identical and understood intent. One single ad can never appeal to two different user intents optimally.
Once ad groups are buying singular intent, it is important that ads are relevant to that intent. This means that every ad in every ad group should be different. For large accounts with may be hundreds of thousands of ad groups this is a very significant operational exercise!. Automation can help here. Ad templates which dynamically insert parameters according to the intent of the ad group can create highly relevant and unique ads at scale. For example a headline of ‘Trains to {destination}’ could be used in thousands of ad groups related to different train destinations. In the London ad group the headline will become ‘Trains to London’ and in the Edinburgh ad group the headline will become ‘Trains to Edinburgh’. Changing the headline to ‘(destination) Trains’ will create ‘London Trains’ and ‘Edinburgh Trains respectively. As such one template can control thousands of ads.
Once you have search query control and efficient ad creation the best ads can only be determined through A/B testing. Running good A/B tests is not a trivial exercise but that is for another article.
Best Landing Pages
Many advertisers build an account structure based around one ad group per landing page. But this method is a ‘supply driven‘ marketing approach. Ad groups built like this will attract users with multiple intent, and ads will not be optimal. Instead a demand driven approach builds ad groups by user intent. Only after the intent of an ad group is determined is the best available landing page then chosen. This does mean that often the same landing page is used with multiple ad groups. As with ads this isn’t optimal. Ideally there would be a unique landing page for each unique user intent.
This does illustrate an important point about search marketing best practice. The very best search marketing companies line up their entire business around their customers. This means they start by gearing their marketing around how people search in a demand driven approach. Next they create dynamic and optimized landing pages that are as unique as their users’ intent. Finally they optimize their conversion funnel, their content, their pricing and even their products and services around their customers. This approach not only delivers a wonderful and seamless experience for users it can also make search marketing an extremely profitable marketing channel. Furthermore by focusing on demand, businesses can find whole new markets to grow into and exploit.
Best Bids
Last but by no means least the best accounts need the best bids. Account structure is often overlooked when considering bidding but it is again of vital importance. The theory of bidding optimisation breaks down if changing bids also changes the intent of search queries that are being acquired.
How can this be? When you have broad or phrase match keywords, bidding up does not just buy you more of the same traffic, it also buys poorer quality search queries that your lower ad rank was not able to acquire before. Similarly, bidding down doesn’t necessarily buy you less of the same traffic. The search queries you were buying may start triggering in other ad groups at a higher price instead. If search queries hop from one ad group to another with bidding changes then bidding optimisation will always be below par. Accounts with search query control have the right foundation so performance per ad group is stable and bids can be set correctly.
Advertisers in different markets may express their marketing goals in any number of ways. One may seek to maximize sales for a given ROI, another to maximize leads for a given CPA or maximize clicks for a given budget. Whilst it is true some goals are better than others, from an optimization stand-point there is a common error we see again and again. That error is optimization that is based on average performance rather than marginal performance. Marginal performance can be a difficult topic to understand, so if you want to know more take a look at our Introduction to Bidding article
In recent months Google’s own bidding methods have become particularly effective. Not only are the bidding algorithms excellent, they are also able to access data they don’t make available to advertisers (e.g. previous searches) and bid on dimensions not biddable to advertisers (e.g. search query level and micro audiences). This, in our opinion, makes the days of the independent bidding algorithm numbered! However, there is often important data Google has no visibility on and significant bidding gains can be made when that data can be put to use. Such data includes product margin, availability changes, pricing and price competitiveness changes and seasonality. In our case, our bidding efforts are now put towards a hybrid approach where we set very dynamic bidding goals for Google at the ad group level. Those bidding goals vary based on metrics Google is not able to determine for themselves, and in this way we are achieving better performance than either they are us can achieve independently.
Conclusion
Search marketing is a mature marketing channel these days with many qualified practitioners. It is rare to find glaring holes in marketing accounts. However as we seek excellence in each of the five key areas (keywords, accounts structure, ads, landing pages and bids) there are still significant gains to be found. We have only briefly touched on the five key areas of search marketing here; there is much more than can be said, and in other articles we can go into more of the details.
We are building our ATLAS product based on these best practices which have been honed through years of hard graft and experimentation in search with a team which has decades of combined experience across over 100 client companies. We are always learning and welcome your comments and feedback.
