Whether it’s performance-driven advertising, awareness or engagement you’re seeking, Facebook undoubtedly offers the best targeting options of all social platforms. The ability to tap into a plethora of your own first-party data, use it for predictive modeling and pair it with third-party data makes Facebook a go-to platform for just about every brand. And as evidenced by some recent releases, Facebook is set on becoming even more widely adopted.
In that vein, there are some exciting new audience targeting features I’d like to share with you, as well as a few lesser-known options driving serious ROI for brands across all verticals. We’ll start with some new targeting for Custom Audiences, then explore other new targeting features and close with some features that have been around for some time but haven’t been widely adopted by advertisers.
New targeting for Custom Audiences
Advanced Custom Audiences
Recently, Facebook began rolling out updates to the Custom Audiences interface to allow advertisers to build a custom audience using multiple data types at once. On July 12, this feature was rolled out to 100 percent of advertisers. The following data types are now available for matching: email, phone number with country code, first name, last name, ZIP or postal code, city, state, country, date of birth, year of birth, age, gender, mobile advertiser ID and Facebook app user ID.
Prior to this change, Custom Audiences only supported matching for one data type at a time: email address, phone number, mobile advertising ID or Facebook app user ID. With this update, you will be able to improve the match rate (coverage) and performance (accuracy) of your custom audiences even further.
(If you are not already seeing this feature in Ads Manager or Power Editor, I’d suggest reaching out to your Facebook representatives and/or your agency partner.)
The “waterfall” approach of providing more data points in order to match to a verified Facebook user recently took a client’s match rate from 50 percent to 80 percent. Facebook also provides some tips for increasing your match rate. If your brand is uncomfortable providing this much information on your customers to Facebook, then there is also the option of creating the custom audiences in data partners like Acxiom, Experian or Datalogix.
From spending a great deal of time understanding the way Facebook hashes customer data, I can say with confidence that your customer data is secure with Facebook and would suggest taking the waterfall approach directly in the ad platform. You can read more about how Facebook creates custom audiences and uses that data here.
Website Custom Audiences
There are some new and exciting targeting options for Custom Audiences created from visitors to your website. These targeting options include: frequency, time spent, dynamic date, aggregated values and devices.
Any account using the Facebook pixel base code can now further qualify a website Custom Audience based on frequency (i.e., how many times a user has visited a particular page) and device. The other targeting options — time spent, dynamic date and aggregated values — are available by making slight alterations to the Facebook Pixel. Time to rustle up some dev resources!
Engagement on Facebook
This new feature allows you to create custom audiences based on user engagement with your videos and lead ads.
For video, audiences can be defined based on level of engagement, like seconds or percentage of video watched. An audience that watches most or all of your videos is valuable because it shows some level of intent. You can now retarget those intent-driven video viewers and sequentially move them down the funnel.
For lead ads, use the new feature to create audiences who started to fill out the form and didn’t convert or those who did fill out the form for retargeting efforts.
Data sets
Fans of predictive modeling will love this new feature. Data sets allow you to create custom audiences based on a number of different customer value attributes that you supply directly to Facebook.
Previously, customer data needed to be prequalified outside of Facebook, in the form of only uploading email addresses, App IDs and phone numbers separated line by line. Now, you can be a data alchemist within one master customer data file. Mix and match customer value attributes like LTV, AOV, Repeat Purchases and so on to create “super audiences” that are qualified on a number of criteria.
Added bonus: Supply just one master list (Do this every so often to maintain data relevance) to your agency partner or internal team members managing your Facebook advertising efforts. It saves you and your Business Intelligence team some serious time.
Other new targeting options
Interest expansion
Interest targeting typically involves the layering of 10 or more interests that are relevant to your desired campaign objective. It can be time-consuming, and it’s also nearly impossible to think of all relevant interest targets on your own.
Let that Facebook algorithm work for you by enabling the hot-off-the-press Ads Manager feature of expanding interest targets. Even better, the opt-in will only expand interest targets when they are likely to produce conversions.
Advanced matching on the Facebook pixel
I’m sure you already have ideas swirling around in your head from the updates shared here so far — but wait, there’s more!
The advanced matching feature allows you to take customer data you collect on your website (e.g., email addresses, phone numbers) and match it with Facebook users. With this additional data, you can identify more users to retarget on Facebook with your website custom audiences or dynamic ads. To enable advanced matching on the Facebook pixel, you will need to modify the default Facebook pixel code to capture your customer data.
This part gets a bit technical, so read carefully and be sure to share this with your dev team.
You can send one or a combination of the following identifiers for matching in clear text or sha256-hashed format: email, phone number, first name, last name, city/state/ZIP, gender and date of birth. Facebook then uses this information to match pixel events in real time with Facebook users, which can lead to more conversions, better ads optimization and higher reach. Note: the data is completely hashed within the user’s browser before it even reaches Facebook’s servers.
Facebook is reporting that during beta testing, there has been an over-10-percent increase in attributed conversions and a 20-percent increase in reach. Facebook has supplied a useful guide to increasing match rates on the Facebook pixel, as well as a developer-friendly guide to getting started.
The lesser-known targeting options
Third-party partner segments (and custom segments)
Facebook partners with a number of data providers like Acxiom, Datalogix and Experian to provide off-the-shelf user segments available for targeting on Facebook. That part you likely knew. However, there are dozens (make that hundreds) of other segments these data partners have that can be ported into your Facebook advertiser account for use.
If you are working with an agency that has a managed partner relationship with Facebook, these audiences can likely be made available free of charge. Otherwise, work directly with preferred data partners to enable these audiences at a cost structure that varies by provider.
Pro tip: Match segments of your customer database against data providers’ user segments. Then you’ll know exactly which segments to use in your targeting. This is a “know before you go” approach that I would definitely recommend. Additionally, most data providers will be able to create custom segments for your brand that can be pushed through your Facebook ad account for ad targeting.
Custom clusters
These are audience targeting groups that Facebook can create for you, based on insights you supply around a desired target audience. Please note, this feature is typically reserved for advertisers working with dedicated Facebook representatives, or for managed partner agencies.
Here are a couple of examples of custom clusters: a custom audience target of Page Admins for B2B Businesses with over 200,000 Facebook followers; a custom audience of close friends and relatives of women who have recently changed their relationship status to engaged.
Since you’re defining the criteria based on Facebook matching users on the back end, they can provide high-value and highly targeted audiences.
Data management platforms
In case you couldn’t tell, I really, really love data. In my experience, thinking like a data scientist pays dividends in a successful Facebook advertising program. Another somewhat unknown targeting possibility to take advantage of involves pushing audiences from your data management platform (DMP) into Facebook through the API.
I like to think of DMPs like a massive file cabinet of customer data and customer prospect data. Most DMPs will even allow for the creation of lookalike and act-alike models within the platform.
Final thoughts
Given that Facebook is all about people-based marketing, finding the very best of the best audiences to target is essential. As social advertisers, we don’t have the benefit of known intent, as a search marketer may have. The name of the game is smart segmentation and modeling of audiences most likely to take the desired action through our “push” marketing efforts.
What are some other ways you’ve found to produce high quality targets on Facebook? Please share!
Comments