Overview
☝ Tracking doesn’t impact performance.
Keep in mind that while tracking is oftentimes important to understanding the value that influencer marketing is providing to your product or brand, it doesn’t impact the underlying performance (and can oftentimes hurt performance if too much tracking is used).
For campaigns with objectives of traffic or conversions, you’ll want to track the impact of creator content on engagement with your site, app, and/or product. This doc describes various options tracking options for your influencer campaign and provides some best practices and considerations for tracking lower funnel campaign performance.
Tracking Options
For campaigns looking track lower funnel activity such as conversion, purchases, signups or other onsite engagement, the section below will walk you through different options to measure onsite engagement.
Terminology
Below are some terms we’ll use when talking about different tracking methods.
Precision (sensitivity). How well a method is at determining the relative impact of different variables to a campaign. High precision methods are usually able to isolate the impact of different changes or variables from other variables in the campaign. Here are some examples of how precision plays a role in measuring impact:
Comparing the relative impact of two creators. If creator A generated 20 conversions and creator B generated 200 conversions, a high precision measurement method could determine that creator B generated 10x the number of conversions as creator A.
Understanding the overall impact of changes. Say you double your creator campaign spend in a given month and that leads to a 70% increase in conversions coming from creators. A higher precision a measurement method is, the closer to the real 70% life the method will report.
Recall (coverage). How well a method captures the total magnitude of impact of different variables in a campaign. A high recall method will mean that you’ll be able to directly attribute most of the onsite engagement to a particular variable. Low recall methods usually require applying a multiplier to the impact in order to account for unattributable onsite engagement. Here are some examples of how recall can play a role in measuring the impact of your campaign:
Understand exactly how many conversions a creator drove. If creator A generated 100 conversions, high recall measurement methods will get close at being able to directly attribute all 100 conversions to the creator, whereas low recall methods will attribute a much lower number to creator A.
Tracking Methods
Below we list different methods of tracking onsite engagement. Note that there’s no perfect method, with each method having different strengths and weaknesses. The list below is sorted by how frequently they’re used.
Custom link
Provide each creator with a custom tracking link that they can post in their link in bio or in their video or post description. These custom links will need onsite engagement tracking in order to measure onsite engagement. The most common way this is done is by using UTM parameters and measuring onsite engagement using Google Analytics.
Strengths:
High precision. If creator A generates 5x more conversion than creator B, you’ll generally see that link click data will reflect the 5x increase accurately.
Full funnel visibility. Depending on your tracking, you’ll be able to see all onsite events that lead up to the conversion event as well, including users landing on your webpage / app, and all the activity that users do afterwards.
Minimal setup work. Custom links are generally very quick to set up. The only engineering work required is making sure that different onsite events are triggering the correctly event logging so that you can measure each event.
Weaknesses:
Low recall. Most of the onsite engagement that’s driven by your 1stCollab campaign will not be directly attributable to the creator that drove the conversion. In our testing, we’ve seen that only about 5-40% of total conversions are directly attributable to the link click. Some factors that impact recall include:
Simplicity of the link. For example, users will usually find it easier to go straight to 1stcollab.com or Google 1stCollab than navigating to a creator’s link-in-bio, and then hunting for the link in a creator’s link-in-bio.
Value in clicking the link. Users are much more likely to click links that offer them special discounts or other value.
Device privacy settings. Depending on the user’s device and privacy settings, cookies might be blocked or link parameters might be striped, which will prevent onsite engagement from being tied to a link click.
Platform and Formats. Clickthrough rates on content are much higher on YouTube long-form content where you can directly place a link in the video description vs. on TikTok, where users can only expose one link in their link-in-bio.
Promo code
Provide each creator with a custom promo code that can be applied by users for certain discounts.
Strengths:
High precision. Similar to link tracking, usage in promo codes is usually good at measuring the relative impact of different campaign variables.
High recall. We’ve seen that promo codes tend to capture around 50-90% of all total conversions. Some aspects that impact promo code usage include:
Value of the promo code. In general users are more likely to use or remember promo codes when they deliver major benefits or offer large discounts to them.
Promo code prominence in content. There are many places creators can feature promo codes: in the content description, displaying it on screen in the content, or talking about the promo code in the video. Reminding users about the promo code in more places gives the codes a higher chance of being used, but might decrease overall content engagement and reach.
Weaknesses:
Doesn’t work for all products. In particular, it’s hard to make promo codes work for free products, where a code won’t provide users any benefit.
Additional cost. Promo code usage often means that users will get discounts on products, which means decreased revenue for the business.
Potential difficulty in implementing promo codes. Depending on your billing system, it might take some work to implement promo codes in your product.
Cannot capture upper funnel engagement. Promo codes only capture engagement at the time of checkout / purchase. So you won’t be able to use promo codes to understand upper funnel engagement such as site visits, landing page engagement, or how many users entered a checkout flow.
Time base changes
Time based or time series analysis involves matching up timelines around when engagement ramps up and down in your influencer campaigns and mapping that to overall site engagement. For example, if you ramp up a campaign and notice large amount of engagement on the content from the campaign, and at the same time see a significant influencer of traffic to your website, you can be fairly confident that the campaign is what drove the traffic.
Strengths:
No implementation. There’s nothing you need to set up in order to do this tracking.
Weaknesses:
Not effective when influencer marketing doesn’t drive a significant portion of overall traffic. In other words, influencer marketing isn’t a major source of overall site traffic (>10% of overall traffic), it becomes more and more difficult to isolate the impact of an influencer campaign, especially when there are so many other variables that impact traffic or lower funnel engagement.
Surveys
User surveys are a good way to get both qualitative and quantitive data about users. In particular, these surveys might ask users questions like: “How did you hear about us?” or “What led you to sign up for our service?”, which can provide an indicator of the impact of influencer marketing.
Strengths:
High recall. Because surveys an indicator of the % of the population that does X, surveys are good for accessing total population metrics. For example, if 10% of all signups answer that they heard about your service through YouTube, that’s likely a good indicator that either influencer marketing and/or organic YouTube ads is a significant driver of signups.
Flexible. Because surveys can be triggered at any time and ask any assortment of questions, you can measure the impact of influencers in various ways and their impact of different aspects of your product (including not just engagement but sentiment).
Weaknesses:
Cost to user experience to accurately measure. High precision surveys often require a large volume of survey results. That might mean triggering surveys for lots of users, which can be disruptive to the product experience.
Product investment. High quality surveys take time to craft and then implement, especially if you want high fidelity answers.
Variable precision. People might not always accurately remember things such as where they first heard about a product, or even which influencer motivated them to check out a product.
Analysis time needed. There’s often quite a bit of analysis work to understand survey results, especially to make sure that as much bias is removed from survey results as possible.
A/B Testing
A/B testing is the gold standard for measuring the impact of different product or marketing initiatives. There exists two types of A/B testing for influencers:
Creator side A/B testing. This means segmenting a group of creators into different populations and giving them different briefs or campaigns to understand the impact of those changes. 1stCollab native supports this form of testing with ad groups, and is very simple to set up. This section won’t focus on creator side A/B testing.
User side A/B testing. This is far more challenging to set up, but can give you very precision and recall when correctly set up. This testing means dividing up user populations in order to understand the impact of influencer marketing. This section will focus in this form of A/B testing.
It is near almost impossible to do rigorous user side A/B testing on influencer campaigns, but it’s possible to approximate such test. As an example, a global brand might decide to launch influencer marketing campaigns in only 50% of their major countries, and observe the growth trajectory of those countries vs. the 50% of countries where influencer marketing was not launched.
Strengths:
High precision and high recall. Assuming your control and treatment groups are similar, you should be able to understand, with a high degree of confidence, the exact impact of an influencer marketing campaign on site engagement.
Weaknesses:
Very complex to implement well. There are many factors that will determine whether an A/B test is set up well (with many of those factors being specific to you brand or product). As a result, we don’t recommend trying to set up a user side A/B testing unless you’re experienced with setting up and running A/B tests.
Can’t be used to understand the impact of individual creators. Unfortunately, we’ve never seen A/B tests be able to measure the impact of individual creators (unless you’re running a campaign with very few, very large creators).
Determining ROI for your Campaign
☝ Improved tracking is iterative. It’s not advised that you set up sophisticated tracking if you’re in the early stages of running influencer marketing. Good tracking takes time to set up, and it might be worth the effort, especially if you don’t have a solid baseline to indicate how much you want to invest in influencer marketing long term.
As you can see from the above, good tracking allows your to accurately determine the precise impact of your influencer campaigns. The main barrier to having a precise indication of impact is the lose in attribution—not knowing where certain conversions came from.
For influencer campaigns, there are many reasons why attribution will get lost with certain tracking methods. Some of the top reasons include:
Privacy or incorrect attribution. For example, iOS will some times remove tracking parameters from URLs depending on a user’s device settings.
Multiple touch points / going through other channels. There might be multiple channels that result in a visit, engagement, or conversion, especially if you primarily rely on last touch attribution. For example, a user might see an sponsored post from an influencer, but then go to Google to look up the website and visit from there. In this situation, last touch attribution would attribute the visit and future site engagement to Google and give no attribution to the influencer sponsored post.
Time it takes to see a conversion. For example, there might be a lag between when a user first sees a sponsored post and when they ultimately decide to convert, especially for larger purchase decisions. If a user sees a sponsored post highlighting an expensive new coffee machine, the user might be in the “consideration” phase for multiple months before making the purchase.
To combat lost attribution, brands will often stitch together multiple tracking methods to come up with funnel efficiency numbers and use those numbers to come up with multipliers for upper funnel or lower funnel metrics. This methodology is best illustrated with an example.
Example Walkthrough
Let’s say say that you’re running an influencer campaign and you’re looking to target a $10 CPA. You’re spending $25k a month on influencer marketing and here are the actual results:
1M impressions / month. The influencers in the campaign drove 1M total impressions on sponsored content each month.
50k website visitors / month. Of those 1M users who see the sponsored content, 50k of those users visit your site each month.
10k conversions / month. And of those 50k site visitors, 10k of those users convert.
So overall, you’re influencer program is exceeding your CPA goes. However, you’re not able to immediately observe the above metrics. Here’s a sample set of steps you can get to to stitch together the above.
Step 1: Get baseline metrics
Before implementing any specific tracking for your influencer program, it’s useful to first verify that your campaign will at least be able to achieve the necessary upper funnel metrics to make your campaign successful.
For example, you might know that your typical new site visitor converts 10% of the time, so in order for your campaign to stay within a $10 CPA, you’ll want to aim for about a $1 CPC.
In your very first campaign, you’ll want to set up link tracking for the campaign. When launching your campaign you might first observe the following metrics on a $25k / month budget:
750k impressions. Site engagement is easy to measure, especially on public engagement metrics.
5k visitors. You’ll measure these based on number of users who engage with your creator referral links.
When first seeing the above metrics, you’ll see a $10 CPC initially and conclude that because you’re 10x above your target CPC numbers, influencer marketing isn’t going to perform well enough to justify the spend. However, the above metrics are actually quite promising. Consider:
A significant portion of engagement on influencer content can happen weeks or even months after content was originally posted. This is especially true on YouTube, which tends to surface older videos frequently, and if your content is getting significant engagement in search, where older videos tend to show.
Observed link clicks often significantly undercount actual site traffic driven. And that’s because oftentimes, most users will just Google for your product rather than finding a link in a creator’s link in bio.
So seeing the above metrics should encourage you to iterate and set up other forms of measurement. If you saw significantly worse results though (e.g. 50k impressions and 500 link clicks), that’ll likely be an indication that should rethink your influencer strategy before investing in further measurement techniques.
Step 2: Set up lower funnel measurement
In month 2, after confirming that upper funnel metrics are within a reasonable range, it’s beneficial to now set up ways to track lower funnel metrics.
As described above, one popular way of doing this is with promo codes, which can help solve many of the attribution problems that link tracking has. Let’s say you set up a promo code to track conversions by giving users who use the promo code a 10% discount off their first month. You also set up conversion tracking through link clicks as well.
You’re metrics might end up looking something like this in your second month:
1M impressions. This increases as you earn more impressions from your first month’s content and impressions from new content from this month.
6.5k visitors. Similarly increases because of engagement happening across multiple months.
1300 conversions / month, measured by link clicks. These are conversions as measured through direct link click attribution. In other words, people who converted through going through a creators link-in-bio link.
13k conversion, as measured by promo code usage. These are conversions, as measured based on promo code usage, which tends to capture a much higher portion of conversions when compared with link click based attribution. Note that turning on a promo code might increase conversions above the long term baseline, since you’re offering a special deal that might in the short term drive more conversions, but lead to lower retention.
You’re big learning from this step is that link based conversion tracking is capturing, at best, 10% of the total conversions that your influencer program is driving. This is an important learning, as it now allows you to apply a multiplier on top of numbers you measure using click based conversions to estimate the impact of your influencer program.
In fact, moving ahead, you can turn off promo code based attribution once you gain confidence around what the correct multiplier should be for click based attribution (though we recommend you confirm the multiplier using other attribution methods and revisit these multipliers on a regular basis). This means if you run you continue running you campaign and notice that each month, you’re able to measure 1k conversions based on link tracking, you can ballpark the total conversion number at around 10k.
Step 3: Optimize your conversion funnel
Once you have gain confidence in your ability to measure events across the conversion funnel, it’s time to begin optimizing your metrics. Given where you’re seeing drop off in the conversion funnel, here are some tips for where you can start to optimize.
Improving upper funnel engagement
This usually means you’re looking to improve how many views or how much engagement you’re getting overall across the creators in you program. Success is usually measured here by CPMs (cost per 1000 impressions) or CPE (cost per engagement). Some techniques for improving these metrics include:
Improving the creative brief. As you’re reviewing creators who are going live, what stands out about the content that performs well vs. the content that under performs? How can you tweak the brief such that you get more of the content that matches the high performers? What other ideas do you have for interesting creative creators can make?
Testing out new creator sets. It’s always worth exploring new types of creators to see how they talk about your product or whether their audiences will resonate with your product offering.
Doubling down on successful creators. We’ll take care of this part for you by automatically signing on successful creators to longer term partnerships. Over time, because we’re continuously refining and improving the set of creators we’re working with, you should see CPM and CPE naturally trend down as your campaign continues to run.
Improving traffic
If your upper funnel metrics, such as CPM or CPE are great, but your cost per click (CPC) is trending high, here are some ways to improve CPC:
Iterating on creative. In particular this might means clarify the talking points of your product or the features creators are showcasing. The overall creative creators make might be engaging, but if users aren’t clicking on their links to learn more, it’s usually a sign that creators are underselling the product
Iterating on the call to action. While it’s a short part of the brief, it’s worth pay particular attention to how creators provide the call to action. Are they mentioning the promo code? Are they mentioning how to find the link? Are they letting users know about limited time offers or sales? Or are they spending too much time making the call to action and the creative feels too much like an ad?
Optimizing format. Different formats and platforms have different clickthrough rates, so if you’re optimizing for CPC, it might be worth biasing more towards particular formats. For instance, clickthrough rates on Dedicated YouTube long-form videos, where you can put a link directly in the description of the video, are usually much higher than TikTok, where users can only include a single link-in-bio on their profile page.
Improving conversion rates
If creator are effective at driving traffic to your products landing page or app but those users aren’t converting, it usually means that the landing page experience could be improved or that the visitors themselves might not be very relevant.
Iterating on a landing page experience. Visitors who come in through creators might be very different than visitors coming to view your product through Google or organic ads. For example, if you’re running campaigns on TikTok or IG, you’ll likely notice a large portion of traffic coming from mobile. It’s worth looking at how traffic from influencers might be unique, and what the right landing page experience is for those visitors.
Testing new creator sets. If users come to your landing page, but bounce shortly after, the visitors to your landing page might not find the product relevant. In these cases, it’s worth testing out different sets of creators to try out different audiences to understand if different audiences have different conversion rates.