Tag Archives: Beta

The Beta Debate

To beta or not to beta…that is the question. It is always tempting for search marketers to load up a text ad with the latest ad extensions offered by Google or Yahoo!/Bing to command more real estate on the search engine results page (SERP), but does it pay off in the form of increased ROAS? The answer is sometimes yes, but a lot of times, no. So before you jump on the beta bandwagon, carefully consider your vertical and the KPIs to which you are managing to ensure the beta will actually help you achieve your goals.

When evaluating whether or not a client should participate in a new beta, the media team at IgnitionOne takes into account the following considerations:

  • Vertical – While it makes sense for a travel or retail client to run Google’s new image ad extensions, the beta will not be as impactful for a telecomm or banking client. Google users will be interested in seeing images of a luxury hotel or fashion house’s latest line, but not as intrigued by a stock photo man on the phone or a pile of cash. Running banal stock images will not increase CTR or any other return metrics.
  • KPI (Key Performance Indicator) – To what metric are you optimizing your client’s account? If you are lucky and the answer is simply site-traffic or awareness, then the world is your oyster (the world of betas that is.) But, if your KPI is ROAS, as it is for most advertisers, then choose your betas carefully. For example, running Google’s communication extension may encourage users to sign up for a newsletter, but it also may discourage them from actually clicking through to the site and potentially making a purchase. Another recent example of a newish product that has the potential to decrease ROAS is Yahoo RAIS Video ads. Because the video does not link through to your site, you are essentially trading potential conversions for video views.
  • Tracking & Reporting Capabilities – If the results are not “trackable”, then they do not exist. While a beta may help drive incremental revenue, if you cannot track the performance and report on it, that revenue (and your effort in setting up the beta) is wasted. For example, enabling call extensions for Yahoo|Bing! campaigns may look great in the SERP and will drive traffic to your call centers, but if your campaigns are not integrated with a call tracking technology, those conversions will be lost. Most times, your technology platform will not have built in support for betas, so you may have to set up workarounds, like dummy tracking.
  • Budget Constraints – Any budget allocated towards testing engine betas should be incremental. Never rely on a beta product, like Google’s DSA (dynamic search ads) or Yahoo|Bing’s in-stream ads to spend your core budget or drive revenue towards your overall goal. Betas should always be tested with a small incremental budget before they are launched across an entire account and expected to boost performance.
  • Client’s Level of Risk Aversion – Now I am scared of many things, including heights, public speaking and tick borne illnesses, but engine betas are not among them. However, many clients prefer to stay with the tried and true and are hesitant to enable betas before they are a proven success. For example, clients with highly particular brand guidelines tend to stay away from betas that have any type of dynamically generated ad-copy, since the predictability of how the brand will appear is low during the beta period. Other clients are not satisfied unless they are on the cutting edge of something, even if the risky behavior costs them performance. Of course, the smartest advertisers fall somewhere in the middle, but always consider your client’s risk aversion when proposing that they test an engine beta.

The bottom line for betas is to proceed with caution. Marketers should carefully consider their vertical, KPI, tracking and reporting capabilities, budget constraints and risk aversion before opting into the latest and greatest engine offering.

Leveraging Rich Ads and Branded Sitelinks in Yahoo!/ Bing

With Yahoo!/ Bing (YaBing) releasing a beta for sitelinks, alongside the release of Rich Ads earlier this year, IgnitionOne investigated how to leverage the budgets of branded campaigns to increase efficiency to marketers’ YaBing account.  The basic premise of leveraging the two products’ budgets to increase efficiency lies in the ability to predict brand click-through rate (CTR).

How Rich Ads Work

Rich Ads run only on exact and phrase match types, can only run on brand terms or extremely relevant non-brand terms and the ad position only shows at a rank of 1. This design of the Rich Ads, along with the fact that no competitors can set up a Rich Ad campaign running on another marketer’s brand terms, indicates that the algorithm which dictates cost-per-click (CPC) for this product is slightly different from the algorithm we see in the perfectly competitive open market (i.e., the non-Rich Ads campaigns’ algorithm). The algorithm in the “open market” determines rank and CPC by looking at Quality Score (QS), which is usually a 10 for branded terms, CTR, as a relevancy proxy, and the next in lines competitors’ bid in comparison to your bid. Due to the Rich Ads’ design we can take out QS as a major determining factor for the CPC you are charged, as well as the next-in-line competitors’ bids. We also know that the algorithm is only predicting CPC as the rank and never changes for Rich Ads. This leaves CTR determining CPC. As we all know, CTR is determined by clicks and impressions.

So how can marketers use this information in tandem with their branded sitelinks strategy?

The Strategy

Since the Rich Ads algorithm primarily gathers information from within its campaign and pulls relatively little information from the market, it has more of a lagged model than the “open market” algorithm. Essentially, it uses past data to predict what will happen to the CTR in the future. The shift between the lows and the highs of brand demand (i.e., impressions) is where this gets interesting.

The lag experienced in CPCs, essentially based on impressions, is about 1-2 weeks (note: this may vary based on the amount of impressions your brand receives). During the period in which your brand has reached the descent from the apex of a high demand period, you will be charged high CPCs for about two weeks into your descent of demand within the Rich Ads campaign without the justification in ROAS (i.e., conversions or AOV decreases). This is when IgnitionOne recommends decreasing your daily budget caps so that your rich ad campaign will, in fact, flight (see Figure 1.0). From here, marketers should reallocate extra budget into branded campaign that has sitelinks, to ensure that all exact, phrase and broad match terms come in a rank of 1 (See Figure 1.1). This is important as the “open market” algorithm will adjust more quickly due to your competitors pulling out of the market/lowering their bids during periods of low demand. The algorithm will adjust more quickly also due to your ability to change more keywords’ average ranks to a 1, thus bolstering CTR, which will in turn lower your CPC. Once you see the Rich Ads’ CPCs coming down to where they should be, your brand should increase the budget to where it will not flight further.

Best Practices:

-Identify if your CTR follows this same pattern during extreme changes in brand demand.

-Watch your Rich Ads campaign to see how it performs during fluctuations in demand. You will need to understand your brand’s CPC lag time.

-Set keyword bids high in the Rich Ads campaigns. You will need to manage spend in your Rich Ads by daily budget rather than CPC, as there are no competitors bidding on these terms and no dispute for rank.

-Set the Rich Ad campaign serving setting to “accelerated.”

-Do not turn off your Rich Ads campaign during the period in which your CPCs will not justify the ROAS. You will need the Rich Ads algorithm to collect data in order to get CPCs where they should be.

-Ensure your branded campaign has enough daily budget to take over the Rich Ads campaigns’ daily spend when you switch over.

NOTE: How to predict changes in CTR

Impressions and clicks rise in periods of high demand for your product or brand. However, it is common that in periods of extremely high demand, clicks are able to maintain at the same rate as the demand (i.e., impressions). Due to this effect, brands typically can experience drops in CTR. Since CTR is the main dictator of our Rich Ads CPC, brands using this product can experience higher CPCs during high demand periods. This is not a huge detriment to the account as conversion rates tend to be fairly high in Rich Ads campaigns and also during times of high demand. Conversely, CTR increases as impressions decrease during a brand’s offseason causing CPCs to decrease in Rich Ads. It is during this time that IgnitionOne would recommend keeping Rich Ads budget caps high.

Figure 1

Figure 2