When Sponsored Collaboration Results "Don't Click"—A Study in the Failures of Casting Design

Let's start with the conclusion. Many cases where sponsored collaboration results fail to "click" as expected can be explained not by the enthusiasm or qualities of the individual influencer, but by flaws in the design on our side—the brand doing the casting. Selecting based on follower count alone, sending requests without fully fleshing out the details, delivering without considering timing, doing nothing after sending—these gaps in the overall setup significantly affect outcomes. In this article, we articulate these four patterns based on our own retrospectives.

Please note that all cases and figures mentioned in this article are anonymized aggregate statistics and do not indicate any specific brand, campaign, or individual. Legal considerations such as stealth marketing regulations are not covered here; this article focuses solely on the design of the overall setup.

Why We're Sharing This Perspective

"We chose the wrong fit, so the results didn't click." When a sponsored collaboration doesn't produce the expected results, this is the explanation most commonly heard on the ground. It's true that if the tone of the feed doesn't align with the brand's worldview, the chances of success decrease. This much is probably already known to anyone involved in the industry.

However, what has become clear to us through our day-to-day retrospectives is that failure doesn't reside solely in "who was chosen," but in "how the arrangement was put together." And selection itself is part of the design. Choosing a good-fit partner is merely the entry point of casting design; it only becomes a complete design when it includes the granularity of the request after selection, the timing of delivery, and the follow-up after sending. In many cases on the ground, it's this "after selection" portion that tends to be neglected.

Four Patterns of Failure

Here we organize, into four categories, the gaps in casting design that we have repeatedly observed in our retrospectives. In each case, the subject is us—the party doing the casting.

First, oversimplified selection criteria. Sometimes we narrow down candidates based solely on follower count or genre tags, without examining the color and tone of their feed or the temperature of their communication style. Even when the numbers are the same, you can't know the degree of alignment in worldview without actually looking at the feed.

How to catch this: Before narrowing down candidates, we actually look at a dozen or so of their recent posts and ask ourselves whether we're judging based only on follower count or genre tags. The process of narrowing down candidates itself is covered in detail in The Success of Casting Is Often Decided at the List Stage.

Second, insufficiently detailed requests. Sometimes we send a request without fully articulating, on our end, exactly what aspect of the brand we want conveyed and in what context. In projects involving multiple departments—account managers, production, sales—it becomes easy for no one to take on the role of "translator," leaving the request document to stand on its own. For the recipient, a request with too little information becomes one where they must find the "reason to feature it" themselves.

How to prevent this: Before sending, we write out, in sentence form, "one thing we want to convey" and "one expression we want to avoid," and align understanding among internal stakeholders. How far to refine the granularity of a request is covered in Even With the Same Product, the Granularity of the Request Changes the Output—A Practical Discussion of Briefing Design.

Third, lack of timing design. Sometimes we decide timing purely based on our own convenience as the sender, without considering seasonality, the individual's posting cycle, or periods when they're juggling other projects. If there's no "reason to post this now" at the moment it arrives, the post tends to get deprioritized.

How to catch this: Before sending, we check recent posting frequency and seasonality, and verify that we haven't compressed the lead time between sending and the desired posting period too tightly.

Fourth, lack of follow-up. Sometimes, after sending, we fail to provide information that would make posting easier—such as tag guidance or the brand's intent—and simply treat the matter as finished once sent. We should be honest that, unlike the other patterns, this one is not backed by quantitative data within our performance records; it is based on hands-on experience in the field. Our sense so far is that the deciding factor isn't reminders or pressure, but whether we've provided information that makes it easier for the poster to take action.

How to prevent this: We operate on a system where, for each project, we check whether we've provided guidance on posting tags or the brand's intent within a set period after sending.

Supporting Evidence

Let us cite one piece of data from the performance records we shared in our inaugural article. In our performance records covering the past 4 years, 44 projects, and 342 individuals, 91% of posts were made even for gifting arrangements with no posting obligation, with 43 of the projects falling within a 90–100% range. Within this, one project alone showed a posting rate of 33% (n=30), pulling down the overall average by roughly 3 percentage points.

The four patterns in this article are not a reinterpretation of this single case as "proof." Rather, it was in reflecting on this case that we first wondered, "Can this really be explained by the quality of selection alone?"—and that question became the starting point for this inquiry. For a detailed analysis of the contributing factors, see What Percentage of Gifting Arrangements With No Posting Obligation Actually Get Posted?.

Not Every Project Needs the Same Level of Design Density

Reading this far, some of you may be thinking, "Isn't it a lot of effort to go to such lengths every single time?" To be clear, we are not arguing that every project requires an equal density of casting design. Projects where the worldview is already aligned and there's little ambiguity in how to deliver the message often reach good results naturally, without needing to brace for it.

Investing in casting design is especially effective for projects prone to mismatched expectations—projects where the business relationship has just begun and mutual assumptions aren't yet aligned, products with complex context to convey, or projects combining genres for the first time. We believe that judging how much effort to invest in which project is itself part of the work of casting design.

What We Confirm as a Final Check in Our Casting Design

Here, mapped to the four patterns above, is a short summary of the items we check before and after sending.

• Check alignment in worldview before looking at numbers
• Articulate the substance of the request before sending
• Decide timing in line with the posting cycle
• Build post-send information provision into the design

Every one of these is something we, as the casting party, can address in advance.

In Closing

When we attribute the cause of failure to the person, that learning can't carry forward to the next project. The reason we first question our own casting design is that it's one of the few variables we can actually control. Rather than settling for "it just wasn't a good fit," reflecting on where in the setup there was room for better design is precisely what improves the precision of our next casting decision. That is the single point we wanted to convey in this article.

Next time, we'll share further insights from another angle, drawing on our performance records.

We'll deliver further updates via our email newsletter. You can sign up below.

Source: Research by STATS INC. (2022–2026, based on our own project records). The figures cited are drawn from previously published aggregate records (compiling only records where posting status was clearly documented; event-invitation-type projects are excluded as they are aggregated separately). These records have been statistically processed so that no individual project, brand, or time period can be identified.