How to Use Micro-Audiences on LinkedIn to Develop Marketing Personas — and Generate Leads

August 22, 2016

This guest post was contributed by AJ Wilcox, who is the founder of, an agency specializing in LinkedIn advertising.

I’m in the fortunate position of getting to train some of the brightest marketers in the world about LinkedIn advertising. Marketers react very positively to my instruction on how to use the LinkedIn platform to develop marketing personas. Since it usually gets such a great reception, I figured it would be valuable to share my approach here today.

In my experience, the vast majority of marketers deploying LinkedIn Ads use it to generate B2B leads. It’s obviously a great reason to be advertising on the network, but there’s an inherent advantage to being able to specify the professional attributes of your audience, and that is testing behavior.

Here’s a practical example how this looks. Let’s say two advertisers are targeting the same population on Linkedin – Human Resources decision makers in enterprise-sized organizations in North America.

Advertiser 1

Advertiser 1 creates a single campaign, whose targeting looks like this:

It’s definitely not bad. It’s targeting prospects whose job function is Human Resources, and whose seniorities are anything manager and above, at companies with more than 1,000 employees.

Now ask yourself what further actions you would take if this campaign performs really well, or even poorly? The way this campaign is set up, there are not many actions you can take to optimize or isolate performance success factors.

Advertiser 2

Now consider Advertiser 2 who set up the following four campaigns:


Notice that these four campaigns are exactly the same audience that Advertiser 1 targeted, except these four campaigns from Advertiser 2 break each of the seniorities into its own campaign.

Advertiser 2 now has four separate buckets, each measuring metrics such as impressions, clicks, spend, and engagement rate.

Also, notice how if you add up all the various segments, they equal 409,000, while Advertiser 1’s campaign only adds up to 375,000. That’s due to a professional’s ability to list multiple current positions, so some profiles will have multiple seniorities. Can you glean any audience insights by understanding that 9 percent of your audience has multiple seniorities associated to them?


Let’s say both advertisers start using LinkedIn’s Sponsored Content with a single offer – a whitepaper offering advice on how HR professionals can recruit more effectively than their competition.

Advertiser 1 had a click through rate of .33 percent and a conversion rate of 5 percent. Advertiser 2 had the following performance by microsegment:


Sponsored Content CTR

Sponsored Content Conversion Rate













Because we split up that broad audience by their seniorities, now we get insights about what level within the organization our content most resonates with. Look specifically at the CTR results, which communicate how engaging our ads are.

If Director seniority HR folks are the most engaged, we’re going to achieve lower costs per click, and thus, lower cost per lead/opportunity with that audience. Operating under the scenario of limited budgets (the reality for the vast majority of us), I’d much rather spend the majority of my ad dollars on that audience. (I’d also like to have my content team produce content that is aimed directly at this engaged audience).

In this scenario, Advertiser 1 will keep creating the same types of content and ads, having both hits and misses.

On the other hand, Advertiser 2 is going to find out which of her audience segments her company’s content produces the strongest engagement. Next, she can create content or tailor messaging to the individual segments that are most productive to her goals. She is now going to see massive advantages, and glean insights that competitors didn’t even know were available to them.

Where to Start with Targeting LinkedIn Sponsored Content?

So, where do you start breaking out your audiences? This depends on what insights would be the most valuable to you. In the above example, we broke out by seniority, which is one of my favorite ways to compare, but I also really enjoy slicing and dicing my testing audiences by:

  • Geography – Might your prospects in San Francisco behave differently than those in Omaha?
  • Company Size – Director seniority at a small company is extremely different than at a Fortune 500 company. Use this filter to tease out your sales viability to large and small companies.
  • Job Function – Have you identified that HR, Finance, and IT are all involved in the decision to buy your product/service? Break them up and track which is the most effective for targeting.

I recommend running audiences for Sponsored Content that are between 20,000-50,000 members. When I see audiences larger than 100,000, I immediately identify them as candidates for breaking out into more informative groups.

Which of Your Audience Segments Wins?

So, let’s say you’ve taken my word on this and ran a smaller audience split test. Now how do you interpret the results?

I view engagement and conversion as being two entirely different hurdles that must be jumped over by every campaign/content combination. Follow this logic to evaluate:

  1. Engagement. The first hurdle you encounter when advertising on LinkedIn (as well as other social media platforms) is getting someone to click on your ad. Consider which group has the highest CTR or engagement rates. Use engagement rate when social signals are valuable for your goals. Engagement rate can tell you that your content is speaking to pain points felt more keenly by that group, and they’re engaging with the content at an elevated rate. This metric tells you about interaction with ad copy, but for the performance marketer, it doesn’t tell the whole story.
  2. Conversion. The second hurdle you encounter in advertising is getting prospects to convert once they’ve gotten to the landing page. Ask yourself which group has the highest conversion rate. 

A word of warning about small data sets – If you only spend $60 on a campaign, the differences in ads will not be statistically valid. Make sure to collect enough data that you can call the test in confidence. If you’re not certain, search for an A/B test calculator–there are many, and they’re all based on the same math.

For me, I generally like to see around 100 clicks on an audience before I break out the calculator, but I’ll call it early if there are huge extremes (no conversions after 40 clicks, or 10 conversions on 40 clicks).

How Long Should You Run Your Test? 

Keep in mind that the volume of data is important here. It's more important than the time frame over which the test is running. As long as you gather the engagement and conversion data above, it doesn't matter whether your test runs over two days or over four months. Once you have the data and your analysis, then you can move onto to forming your next test. 


Depending on the results of your testing, you will gain insights into many factors that will improve your marketing across all channels. Here are a few examples of the insights you might gain:

  • Which audiences should be bid up to capture highly-profitable traffic in higher quantity?
  • Which audiences should be bid down or paused entirely?
  • Which audiences need new landing pages, messaging, or content?
  • How else can we profitably target our most valuable personas?

On no other platform can you get the same B2B granularity of testing your marketing personas, so of course, continue to advertise on LinkedIn. Take advantage of micro-audiences. This will improve your efforts across all marketing channels!

For more information on how to get the most out of LinkedIn advertising, download "Laser Focus: 10 Ways to Optimize Your LinkedIn Sponsored Content."