Is Your Bid Op Tool Worth The Cost?

By Paul Benson
Thousands of online advertisers in the U.S. rely on bid optimization tools (BOTs) to make their paid search campaigns more efficient.  These software “solutions” or “platforms” which typically leverage rule-based or algorithmic bidding, claim the upper hand against manual bidding strategies.  While it’s certainly true that in some cases these tools vastly outperform manual bidding, the challenge for advertisers is determining whether a BOT will prove cost-effective for their particular campaigns.

BOTs are perceived as superior to manual bidding because they are able to leverage historical data and automatically update bids based on the day of week, time of day, and other factors.  They also save time as advertisers no longer need to analyze performance and make bid changes on a relatively routine basis.  Unfortunately, just as BOTs create many efficiencies, they also generate significant costs.

On average, advertisers pay 5% of spend for BOTs.  For an advertiser spending $100,000 a month, this equates to $60,000 per year.  Some providers also charge a one-time implementation or launch fee up to $10,000.  At this cost, a company could hire a new specialist simply to handle bidding for the account..  It is also true that BOTs require integration with existing platforms, which may cause on-going frustration among advertisers looking for a truly “automated” solution.  This frustration is intensified when advertisers learn that BOTs will require consistent manual oversight, and sometimes intervention.  On a past account audit it was determined that the advertiser’s automated tool had spent nearly $6,000 in two months on a single keyword that generated an ROAS of just 27% (significantly below the ROAS goal).  Neither the tool, nor the keyword, was a new addition to the campaign, which made this gross oversight on the part of the BOT even more perplexing.

For this same advertiser, we noticed over the course of several days that only 4% of converting terms, on average, received a daily bid change, and 88% of converting keywords didn’t receive a single bid change throughout the monitored time period.  Furthermore, keywords that comprised nearly 80% of all conversions were displayed in a position of 1.7 or better.  Nearly 94% of these terms had a CPA at or below goal, signifying that little optimization could be done from a bidding standpoint to improve performance on these terms.  After analyzing the number of bid changes, and the size of each bid change, it was concluded that this particular bid tool saved the client between $500 and $1,000 per month; the bid tool alone cost more than four times that amount to run.
Innovations within the primary advertising platforms, AdWords and AdCenter, make manual bidding all the more attractive.   For example, marketers formerly relied on BOTs to factor in day of week and time of day when making bid changes.  However, now Google AdWords’ Ad Scheduling functionality allows advertisers to adjust bids (up or down) during specified days or times of day.  AdWords also gives insight into performance by time of day or day of week (as long as you have a Google Conversion Pixel in place).  As a result, advertisers can leverage day parting to effectively change bids by time of day and day of week, just like a BOT can.  All of this functionality is available for free directly within the AdWords interface; you don’t need to access, pay for, or become familiar with a 3rd party interface.

Despite the aforementioned challenges, there are certainly scenarios where a BOT is still a valuable and unrivaled tool.  Ultimately, the decision to implement a BOT should be evaluated on a case by case basis.  Fortunately, there are three criteria that will help point you in the right direction.  First and foremost, consider your industry.  Advertisers in the retail space, especially those with thousands of SKUs, are more likely to benefit from a BOT.  This is true simply because a BOT’s value increases as the number of keywords increases.  Sophisticated BOTS can accurately determine the right CPC based on not only historical performance of a particular keyword, but also on historical performance of related keywords.  This insight, although small on an individual keyword basis, can add up to thousands of dollars across an account.

The second consideration can be referred to as the PPC Gini Coefficient.  A variant from its economic definition, the PPC Gini Coefficient can be calculated by dividing the number of converting keywords that are meeting or exceeding your goal and are in an average position better than 2 (1.9, 1.8, etc.) by the number of total converting keywords.  The higher your PPC Gini coefficient, the less opportunity there is for bid optimizations, and therefore the less useful a BOT will be.  A coefficient of 70% is considered high, while 50% is considered relatively normal.  It’s possible that your campaigns contain keywords that are exceeding your goal, but are in a position below 2 (2.1, 2.2, etc.) simply because the CPCs haven’t been properly adjusted.  If this is the case, these terms should be removed from the calculation.

Finally, if you’re currently using a BOT, you should also consider the frequency and size of bid changes.  This data, along with click volume, can give you a reasonable sense for how much money the BOT is saving you every month.  Keep in mind, however, that every time a BOT decreases a CPC to gain efficiency, it may be also sacrificing conversions.  It’s also important to understand whether the bid changes are more commonly max CPC increases or decreases.  Surprisingly, out of all the bid changes implemented for the advertiser mentioned earlier, 94% were decreases in max CPCs.  Therefore, in this particular case, only cost savings were considered; potential revenue generated from increased CPCs was ignored.

If you’re still struggling with the decision of whether to implement a BOT, consider contacting one of the providers to request an estimate.  Marin Software has an internal tool that estimates your increased ROAS if you were to adopt their software.  While these estimates are commonly ‘optimistic,’ this additional information may aid in your decision.  It’s difficult to test a BOT, since significant upfront work is required to integrate the software with your account.  Therefore, you should collaborate with an experienced SEM expert or agency to conduct a thorough analysis based on historical campaign performance.  If you’re already using a BOT, don’t fall victim to the sunk cost fallacy.  Carefully evaluating the efficacy of your bid tool could save you thousands.

Google Introduces AdWords For Video

Earlier this spring, Google expanded their paid search advertising capabilities by launching their new video advertising platform “AdWords for Video.”  The program’s launch immediately turned video content advertising into an affordable, targetable, and measurable medium.  AdWords for Video operates under a pay-per-view model where the advertiser is only charged when users have watched their video in entirety, or for thirty seconds—whichever is shorter.
In terms of set-up, all an advertiser needs to get started is a YouTube Account.  The AdWords for Video platform is programed so that videos from any linked YouTube Account can be pulled directly into new ads in the account.

AdWords for Video regulations allow for four different ad formats:

  • In-Search – As a featured ad above the YouTube search results (similar to the ad locations for search network text ads).  This ad format was formerly known as “promoted videos” on YouTube.
  • In-Slate – As a an uninterrupted featured video that plays before targeted content
  • In-Display – As a suggestion to the right of a targeted YouTube on the video watch page
  • In-Stream – As a ‘skippable’ video that plays before targeted content

Upon learning about the new AdWords for Video platform, our team wondered how it differed from having traditional video ads on the display network.  In our experience, so far there are several key differences.  First, as described above, AdWords for Video allows for cost-per-view bidding.  Display network bidding only allows for CPC, CPM or conversion focused bidding.  Second, the display network allows for non-TrueView format videos to be incorporated into click-to-play or in-stream ads.  Display network video ads also currently allow for placements on YouTube, but Google has already hinted that these ad formats will be phased out to give way for AdWords for Video.  Third, the ability to target ads is different in AdWords for Video.  Instead of having Ad Groups, AdWords for Video organizes ads into “targeting groups” set at the campaign level.  Targeting groups allow for demographic, topic, interest, placement, remarketing, contextual keyword, and search keyword inclusions and exclusions.  Finally, in AdWords for Video, one video ad can be applied to multiple ad formats.  On the display network, each video has unique content and format.

So what type of benefits can AdWords for Video bring search marketers?  There are certainly long term benefits associated with the efficiencies AdWords for Video brings to video content management and optimization, but there is also a more immediate advantage.   Because AdWords for Video is so new, it is unsaturated.  The paid video advertising market is just beginning to develop, and consequently cost-per-view prices in many industries are still available for just a couple of cents.  Also, the cost-per-view bidding format, in of itself, precipitates a major advantage.  Since the advertiser is only charged when a viewer watches their entire video, there is the potential for a lot of “free branding.”  In other words, a partial video views may not be a bad thing.  In fact, if your ads are engineered well and you can get your message across in the first five to ten seconds of your video, you are likely to reach a significant amount of customers for free.

Implementing Segmented ROI Analysis

By Mark Casali
To succeed in online marketing, an agency must be committed to adaptation, innovation, and, of course, optimization.  The industry seems to produce a never ending array of methods to improve and refine account performance, but as search marketers, our very own measure of success remains a stagnant and undeveloped concept.  Return on Investment, or “ROI,” is typically applied as a blanket-metric, and only assessed at the account level.  This evaluation of ROI is distorting goals and leaving perhaps the most obvious optimization opportunities unexposed. In an industry that survives on advancements, it is essential that we begin assessing value and success with more comprehensive and revealing measures.

The easiest method of improving the ROI measurement is to move away from an account level assessment.  Regardless of how the marketer ultimately defines ROI, each product/service category should be tracked with an individual return on investment goal.  Consider a sporting goods store that has a mature paid search marketing account to promote baseball, football, and soccer ball sales.  The store incurs $1,000 of engine spend on each of the three product categories.  Revenue earned by the paid search account is $2,500, $2,000, and $800, for baseball, football, and soccer balls, respectively.  The store has identified an ROI goal of 1.5 as their break-even threshold; this goal includes variables such as product cost, growth strategies, and lifetime value.  Account level ROI is calculated as $5,300 of total revenue divided by a total $3,000 of engine spend, for a return of 1.8.

Most agencies will look at this data and determine that the account is healthy with an ROI well over their goal of 1.5.  Additional optimization techniques could be performed within the account to improve click through rates, strengthen conversion rates, and refine cost per click bids.  These strategies may well benefit account performance, however, there is a more immediate problem that will be exposed by refining our ROI calculation.  If category level ROI metrics are analyzed, the marketer will determine that returns equal 2.5, 2.0, and 0.8 for baseballs, footballs, and soccer balls, respectively.  With an ROI less than 1.5, the client is earning additional incremental revenue in the soccer balls channel, but they are detracting from their overall profitability.  In fact, the soccer balls product channel would have to nearly double revenue to reach the desired ROI goal.  It is unlikely that traditional optimization techniques bridge this gap, especially in a mature campaign.  In reality, traditional optimization in this account is insignificant.  Segmented ROI analysis has revealed a far more pressing performance issue, and until the soccer ball campaign is suspended, the account, from a profitability standpoint, will significantly underachieve.

This example is admittedly basic, but the underlying principle of identifying and striving towards multiple ROI goals is a powerful optimization technique.  The sporting goods store has a simple product offering of three different types of equipment, but actual companies market countless amounts of products and services.  Logically, the marketing campaigns promoting these products and services will achieve various levels of success, and some will even fail.  So, with such drastic levels of performance across an account, it seems counterintuitive to optimize around a one-size-fits-all ROI goal.  A better approach is to calculate separate ROI objectives for each product/service category.  This strategy is very much in line with the way search marketers update bids based on keyword performance.  The marketer would never bid on a keyword with a negative contribution margin; why then would they apply budget to an entire product category that loses money?  This granular approach of assessing ROI on a category/product level is applicable to companies ranging from full scale e-retailers with numerous product lines to institutions of higher education with multiple programs of study.

Once marketers move towards a more segmented ROI analysis, they should also reconsider how they set performance goals.  When establishing an ROI goal, the agency or client should develop customized objectives for each product/service category based on a variety of variables including historical performance, the product costs and gross margins of the related products/services, growth strategies, and the assessed lifetime value of a customer associated with a product/service.  For example, a marketer may determine that a particular campaign drives so many recurring purchases that a return on investment goal under 1.0 could, in fact, be appropriate and representative of long-term profitability.  When establishing ROI goals, the marketer or client should also acknowledge the motives of the campaign.  If the ROI goal is set too high, the profitability of the account will be strong, but the volume and reach of the campaign will be restricted.  Conversely, if the ROI goal is too low, the campaign will see stronger volume but lower margins.

Overall, the limitations of an account level return on investment assessment are becoming more and more problematic for marketers.  Optimization opportunities are lost, and agencies are left striving towards goals that may or may not be advantageous to the client.  To correct this problem, companies should track performance for each product/service category and even identify separate financial goals for each of these categories.  If this level of granularity is already available, the data should be shared with the company’s search agency so the paid search campaigns can be managed more accurately.  The resulting account will offer a deeper level of visibility into performance and help online marketers quickly and easily improve profitability.  This is one of the last pieces of low hanging fruit available to paid search advertisers; don’t let it go to waste.