THE RKGBLOG

Utilizing Business Intelligence in Paid Search Strategy

While shooting for a simple return on ad spend (ROAS) or cost per lead (CPL) target in paid search is a fine way of keeping spend profitable, business intelligence considerations can help to augment the primary KPI in order to optimize for additional goals. These goals may include:

  • New customer acquisition
  • Shifting customer mix (more B2B, less B2C)
  • Lead quality improvement
  • Inventory exhaustion
  • Expanding product selection into new categories
  • CEO mandates – random, reasonable, and everything in between

These are not to be confused with microconversions, or the interaction events that occur outside of the shopping cart which could lead to sales down the road. Microconversions include email signups, catalog requests, and account creation, which can be assigned value for the purpose of properly measuring the impact of paid search within existing ROAS goals. However, microconversions can be utilized to achieve additional goals.

Instead, business intelligence can be used to adjust how the primary efficiency metric is measured, thus impacting bidding and strategy.

Once you’ve established what considerations are important to your business, there are a few different ways you can choose to adjust how you measure the success of your paid search program. To elaborate on these tactics, we use a hypothetical scenario where the primary goal is revenue and the secondary goal is new customer acquisition.

Only Value New Leads/Sales

One option is to only value sales coming in from new customers in your attribution model for bidding. You naturally will still get existing customers this way, however you will only be bidding using the data from new customers. This will inherently shift spend away from areas of your account that are driving mostly existing customers.

One thing to consider here is that by counting fewer orders within your paid search program, you will have to adjust your target to account for the possibility that you will be attributing significantly less sales. Additionally, you may have to account for less data that can be used in bid calculation. In this regard, you will want to ensure that you have a bidding platform that can handle thin data well.

However, unless there are stark contrasts between new customers and existing customers, we do not usually recommend going this route as orders from existing customers are likely incremental when coming through non-brand paid search.

Incorporate Additional Value into New Leads/Sales

Another possibility is that, instead of totally devaluing those orders which don’t meet your initiative, you can add additional value to those orders which do. In our hypothetical scenario, this would mean still valuing those sales which came from existing customers, while baking in added value beyond the strict dollars and cents of the orders for those that come from new customers.

Of course there is the question of how much additional value to add to an order from this target audience. If you are capable of calculating the lifetime value of a new customer, we recommend using some function of this for weighting new customers more heavily.

Something to consider when weighting a conversion type more heavily, however, is that when enough additional value is added to orders of a particular type, the CPA becomes correlated with your newly adjusted ROAS, rather than your original ROAS. This is because if the value added in the adjustment dwarfs the actual order value (because of, say, huge lifetime value), the conversion itself becomes more important than the sales dollars from the order, as shown in the charts below depicting category level performance. This is not a necessarily a problem, but something to be aware of.

Leveraging Existing Analytic Strategies

Once you’ve established what considerations you’ll be targeting and how sales will be attributed, there are several paid search products and analyses which can be utilized under this new lens. Abandoning our hypothetical scenario, we examine a few.

Retargeting Audiences:

Google’s Remarketing Lists for Search Ads (RLSAs) allow for better targeting of desired customer groups by allowing advertisers to show ads only to specified audiences.

Some RLSA specific tactics grouped by sample target business intelligence components include:

New Customers

  • Target people completely unfamiliar with your brand by excluding site visitors
  • Target new customers by excluding past converters

Customer Base (B2B vs. B2C)

  • Create audiences from users who trafficked B2B or B2C-centric categories and pages to target to

Higher Quality Leads

  • Create audiences from users who have been to pages with higher percentages of qualified leads

Geo-Targeting:

In the new world of Enhanced Campaigns, it has become much easier to modify paid search bids based on location. As different areas can vary significantly in how they perform relative to business intelligence initiatives, adjusting bids to account for these differences can help to maximize the success of your program.

We took a look at one advertiser who wants to shift their paid search customer base to be more B2B heavy. Below is a map showing volume of new B2B customers by state, which clearly shows the dominance of New York, California, Texas, Florida, and the Mid-Atlantic states.

However, when we analyzed state performance by percentage of new B2B orders compared to overall order volume, we saw a very different map.

States like California, New York, and Texas were seeing large volumes of new B2B orders because there are simply a larger number of potential B2B customers there. Many states in the Mid-West, however, appear to offer a higher percentage of B2B customers, possibly warranting increased spend in these areas.

Device Performance:

Similar to geo-targeting, Enhanced Campaigns offer the ability to adjust smartphone modifiers appropriately for variance in performance on these devices compared to desktops and tablets. This is vital as the revenue per click from mobile traffic is 69% less than that of desktop traffic, according to RKG’s Q4 Digital Marketing Report.

When measuring the value of these devices, calculating the degree to which they drive customers that meet your business intelligence goals can help to better inform the modifier values. For example, one advertiser focused on new-to-file customers found that non-brand new customer acquisition levels were significantly higher on mobile.

Thus, they may not want to pull back as heavily on smartphones as a simple ROI analysis suggests.

Ad Format Performance:

Different ad formats may also provide different levels of performance in regards to your secondary goals. For example, one advertiser attributes over 30% of all PLA orders to new B2B customers, while that figure is only around 20% for non-brand text ads.

One possible reason for this is that B2B customers know what product they are looking to buy or replace and are quick to order if they see it among the PLAs, while B2C customers may be shopping around more. These differences should be accounted for as you allocate spend across different formats.

Microconversions:

While microconversions are generally considered to be part of the sales process, they can also be used to help achieve additional goals such as new customer acquisition and shifting customer type mix.

In the case of new customer acquisition, new customers are predominately going to be the users signing up for emails and creating new customer accounts. If advertisers have a secondary goal of new customer acquisition and are calculating the value of these microconversions based solely on the expected eventual purchase, there is the possibility that microconversions are being undervalued. Additional value may need to be considered due to the fact that new users are the ones signing up for emails.

Regarding a shift in customer mix (B2B vs. B2C), if there are multiple types of account (i.e. “Business” accounts and “Residential” accounts) then advertisers have an opportunity to value these different types of account creations differently. How much differently will depend largely on differences in average order value (AOV) and lifetime value (LTV) between customer types.

Conclusion

In conclusion, we wanted to highlight a few tactics for valuing and achieving success with regards to business intelligence components, but there are certainly more out there. The main thing to do within any of these tactics suggested is test, test, test! Every business is unique and business intelligence considerations will be unique as well. Find the balance between your primary efficiency target and additional goals, and you will be able to better optimize for overall success in paid search.


  • Ryan Ottino is an SEO Specialist at RKG.
  • Comments
    One Response to “Utilizing Business Intelligence in Paid Search Strategy”
    Trackbacks
    Check out what others are saying...
    1. [...] Utilizing Business Intelligence in Paid Search Strategy, RKG Blog [...]