In this era of diminished growth expectations, companies are demanding greater accountability from their marketers. On the surface, paid search marketing seems like one of the most accountable ways of spending marketing dollars. In fact, digital marketers are increasingly spending a greater portion of their budgets on paid search, almost $40 billion on Google alone in 2012. This total grew by more than 20% year on year.
Managing paid search marketing is a complex process, and it takes a lot of discipline and skill to win the game. The rules are changing fast, competitors respond aggressively, and there’s a flood of Big Data available to analyze. Plus, consumer habits are changing every day.
To meet their goals, digital marketers need to get several things right, including:
- targeting the right consumers who are searching for a solution on the search engines (by selecting the right keywords to bid on);
- placing the right bids on these keywords;
- writing persuasive ad copies;
- make a compelling offering to the visitors who land on your website;
- make it easier to convert or buy; and
- remarketing to those who did not convert.
So, what does it take to be successful? To increase revenue, marketers need to develop a cohesive strategy that is built on a clear understanding of how each piece of this puzzle works and interacts with the other pieces. Smart marketers are now taking a data-driven approach to managing all of these pieces in a systematic and disciplined way. In theory, most believe that leveraging their digital marketing data will help their companies dramatically improve their search marketing results.
Leveraging data in all aspects of paid search marketing, including copy testing, keyword bidding, and organizing the campaign structure is the ideal way to go. But marketers face two critical hurdles in capitalizing on this data: Big Data technology and skilled analysts.
Let’s take a look at one of the most complex pieces of the search marketing puzzle: keyword bidding. This involves placing the right bids on the keywords–neither too high nor too low–at a level that helps them meet or exceed their overall campaign goals. In a typical enterprise, there are multiple campaigns, and there are multiple (several thousand) keywords in each campaign. Overall, the number of keywords can run into millions. Depending on resources available and the size of the account, marketers follow a variety of approaches for optimally bidding on these keywords, ranging from basic to advanced.
The simplest of approaches involves starting with the same bid for every ad group, based on metrics such as click-through rate. More refined methods include using some metrics that tie dollar value of conversion or sale to the keyword bid, at least for those keywords that have had heavy traffic (and hence, reliable, measurable data). The problems with the simpler approaches are three-fold: the individual keywords are washed in a sea of averages, it doesn’t work for a big chunk of keywords, and there’s no real accountability since the bidding decisions aren’t optimal.
The most cutting-edge approaches combine cross-channel data from a variety of data sources, such as click-stream data, keyword semantics, advertising data, and conversion or purchase data. The next step is to determine the right attribution to these keywords and channels using custom attribution models. Once the data is prepared this way, the keywords are scored using predictive models. The scores predict the cost, clicks, and conversions elasticity for various bid levels for each keyword, with a certain level of confidence. This is possible even for those keywords that don’t have enough historical transactional data, by using non-transactional attributes such as similarity, semantics, etc. The optimization algorithms then make the right “allocation” of budgets to these keywords to meet the overall portfolio goals. Once the bids are determined this way, they’re placed in the search engine bidding platforms. Bids can be refreshed as often as needed, and models can be rebuilt based on their shelf life.
For example, marketers can use bid-optimization analytics to unlock the value of leveraging deep data mining practices in pay-per-click search and identify the value of every keyword in a campaign portfolio. Users can then launch new keywords based on data gathered from multiple sources, removing the “guesswork” that marketers have typically turned to in the past.
Keyword bidding is only one piece of the whole puzzle. There are other pieces that benefit from technological innovations and Big Data. Data and technology investments help make analysts more productive, for instance. They free analysts from routine operational details, allowing them to focus on gathering a deep understanding of consumer behaviour and deriving actual insight. Analysts, in turn, must have a detailed understanding of digital marketing, as well as the ability to translate complex data into simple insightful stories.
The data-driven approach, combined with the right technology and the right analysts, more effectively serves the customer by connecting them to relevant products and information. In turn, an organization will maximize its search marketing spend and increase its overall revenue.
Venkat Viswanathan is CEO and founder of LatentView.