Since pay-per-click programs first launched in the mid-1990s, Search Engine Marketing (SEM) has been an effective way for marketers to connect with consumers who are actively seeking out a brand’s services or products. However, as search demand has grown, factors like costs, management complexities and competition have significantly increased. To overcome these challenges, tools and practices that leverage data have been developed to help marketers meet their goals. This post summarizes a few of the ways marketers can use data to improve SEM performance.
Application Programming Interface (API) is a language and message format that computer programmers use to dictate how an application program interacts with operating systems, database management systems (DBMS) or other programs. APIs are an important tool for marketers because data collected by external sources can be used as contextual signals to adjust a campaign’s settings to better serve goals.
For example, let’s say a marketer’s business is negatively impacted by cold, rainy weather. By referencing a marketer-created Google spreadsheet that stores a list of campaigns and their associated locations, marketers can use Adwords to automatically adjust bids based on the weather. Adwords accomplishes this by referencing weather data through the OpenWeatherMap API and, after a few calculations based on a certain set of rules, applies corresponding location bid multipliers to adjust an ad’s rank.
When marketers get creative with relevant APIs, this data can greatly impact bidding and performance. For example, 360i leveraged external APIs to bid based on flight cancellation data in its “Turning Flight Cancellations into Hotel Reservations” campaign for client Red Roof Inn. The winter of 2013 was one of the worst in decades, causing on average 500 cancelled flights and 90,000 travelers to be stranded every day. By leveraging flight cancellation data through APIs, we were able to increase bids on hotel-related queries around troubled airports and drive stranded travelers to book rooms at Red Room Inn. Leveraging this data to influence the SEM campaigns increased Red Roof Inn’s share-of-voice by 650 percent on key travel search queries, increased conversion rate 375 percent and lifted bookings from non-brand terms by 60 percent.
Google AdWords Scripts
While marketers can use external data sources to increase performance, Google AdWords scripts can be used to leverage an account’s data to improve metrics. For example, marketers could use the Bid Testing script to test performance at different automated bids and then optimize to the best results. If a marketer notices that a keyword’s performance does well at a certain rank, the Bid-to-Position script can help the account maintain those positions.
While many scripts are related to automating management to optimize performance, scripts can leverage data to alert marketers when something may be wrong. The Account Anomaly Detector compares impressions, clicks and cost against the same time periods for the previous 26 weeks, and if front-end metrics are not within a definable range of acceptable performance, the script sends an email to alert the marketer.
Monitored Search Behavior
Recognizing what terms searchers are querying on can give marketers insight into consumer behavior that can be leveraged to increase performance. Google Trends indexes search query popularity over a user-designated time, and since the data can be filtered by geo, time and category, it can give marketers high-level insights that are actionable. For example, if a retailer wants to see what brands are gaining in popularity so they can take advantage of that demand, they can use the tool to identify what terms are garnering the most queries. The retailer can then bid on those queries with the hope of driving conversions.
A more sophisticated tool marketers can use to monitor search behavior is Hitwise. A commercial platform that measures online behavior by collecting data directly from ISPs, Hitwise reports visitor behavior that marketers can act on. For example, it can report how a visitor got to the site from search (either through a paid or natural link), what terms drove them, where they were before and after the visit, and much more. Given these insights, marketers can adjust their accounts to appeal to these behaviors through keyword, creative and site changes.
Collected Consumer Behavior
Millions of consumers click on SEM ads daily, and each click provides marketers with data as to what and how they are searching – what device they are using, what their location is, when they are searching, and so on. Marketers have the ability to utilize this information to make informed decisions and adjust paid search tactics. For example, if mobile searches in Atlanta spike on Wednesdays at 2 p.m. and stay steady until 8 p.m., marketers can apply settings that take advantage of this data.
On-site consumer behavior data can also be used to improve SEM performance. For example, landing page testing can reveal what user experience drives the most conversions, and conversion pathing analysis can determine which pages can be optimized to help create a smoother, more effective conversion process.
Deeper analysis that connects why searchers are querying and what their on-site behavior is can unlock additional performance. For example, let’s say a travel marketer recognizes the latency between a search query and a booked vacation is very short as back-to-school season approaches. That marketer may recognize that there is demand for last minute vacations and adjust creative to take advantage of this insight.
Search has been described as the world’s largest ongoing consumer survey. As search behaviors, engine search result pages and SEM continue to evolve and grow in complexity, marketers will continue to face challenges in connecting with searchers and encouraging them to “vote” on their paid links. By wisely implementing tools that leverage data both inside and outside the auction, marketers can help ensure they are driving traffic and performance that serves their goals.
Cover photo via CSCSS