Agency Growth on TikTok (Why It Was Hard)

In the mid-1800s, the California Gold Rush drew thousands of people west with the promise of easy wealth. Most found that while there was gold in the hills, the cost of the tools, the volatility of the land, and the sheer competition made it nearly impossible for the average prospector to build a lasting business. I have seen a similar pattern emerge in the digital space over the last decade. Many agencies rushed into the short-form video market expecting quick wins, only to find that the environment was far more taxing than they anticipated.

Navigating the Volatility of Recommendation Engines

This section breaks down how the shift from social graphs to interest-based algorithms created a high-risk environment for marketing firms. We examine why the unpredictable nature of the “For You” feed made it difficult to provide clients with the consistent, month-over-month growth they expected from their digital investments.

When I first started tracking the performance of the TikTok algorithm, I noticed a significant departure from how other platforms functioned. Most legacy systems relied on a social graph, which means you see content from people you choose to follow. TikTok popularized the interest graph. This system uses a recommendation engine to show users content based on their behavior, regardless of who they follow.

For an agency, this change was a double-edged sword. While a single video could go viral and reach millions, there was no guarantee that the next ten videos would do the same. I managed a project for a boutique fashion brand where one video hit three million views in forty-eight hours. The client was thrilled and immediately increased the budget. However, the next five videos we produced struggled to break one thousand views each.

This “lottery” effect made platform comparison analysis very difficult. In a traditional setting, you can usually predict that if you spend X amount of dollars, you will reach Y number of people. On TikTok, the organic reach comparison between two similar accounts often showed wild discrepancies that had no logical explanation. This volatility is the primary reason why scaling a client’s presence felt like chasing a moving target.

Defining the Recommendation Engine Trap

A recommendation engine is a set of rules used by a platform to suggest content to users. Unlike a “Following” feed, which is chronological or based on friendships, a recommendation engine prioritizes “watch time” and “re-watch rates” above all else. This means your content is constantly competing against every other video on the app for a few seconds of attention.

The reason this hindered agency growth was the lack of “compounding interest.” On other platforms, as your follower count grows, your baseline reach usually grows with it. On TikTok, I observed that even accounts with a million followers could see their views drop to nearly zero if the algorithm decided a specific video didn’t meet its immediate retention benchmarks. This forced agencies into a cycle of constant reinvention, which is not a sustainable business model.

• Platform-native retention signals: These are the data points the app uses to decide if a video is “good.” They include how long someone watches before swiping and if they check the comments.
• Organic reach decay: This refers to how quickly a video stops being shown to new users. On TikTok, the shelf-life of a video was often less than 24 hours unless it hit a specific viral threshold.

The Unsustainable Cost of Creative Velocity

In this section, we analyze the heavy resource demands required to stay relevant on a platform that thrives on “newness.” We look at the data regarding how quickly content loses its effectiveness and why the traditional agency model struggled to keep up with the necessary production volume.

One of the biggest hurdles I encountered was the sheer volume of content required to maintain a presence. In my experience, a standard ad on most platforms can run for several weeks or even months before you see a drop in performance. On TikTok, I saw “creative fatigue” set in within days. Creative fatigue happens when the target audience sees an ad so often that they stop engaging with it, causing the cost-per-click (CPC) to skyrocket.

To combat this, agencies had to adopt a “high-velocity” production strategy. This meant moving from producing four high-quality videos a month to producing four videos a day. For many firms, the math simply didn’t work. The labor costs of filming, editing, and trend-tracking outweighed the management fees they could reasonably charge their clients.

Understanding Placement-Level Performance Metrics

When we look at social channel optimization, we have to look at the “click-through rate” (CTR) at the placement level. A placement is the specific spot where your ad appears, such as between organic videos in the feed. I tracked these metrics across twenty different client accounts over an eighteen-month period.

Interestingly, the data showed that the more “polished” a video looked, the worse it performed. The audience preferred raw, handheld phone footage. This created a paradox for agencies. Clients often pay for high production value, but the platform rewarded “low” production value. I often had to convince executive boards that the expensive camera crew we hired was actually hurting their ROI.

• Creative Fatigue: The point where an audience ignores an ad because it is no longer “fresh.”
• Spark Ads: A native ad format that lets you turn an organic post into an advertisement. While effective, these required a constant stream of new organic content to fuel the paid side.

The Struggle with Audience Demographic Trends

This section examines the difficulty of reaching high-intent buyers within a platform dominated by younger, entertainment-seeking users. We discuss the challenge of aligning client products with the specific habits of the user base and the limitations of the available targeting tools.

A common pain point for the managers I work with is justifying spend when the audience doesn’t match the product. According to research from the Reuters Institute, the primary reason users visited TikTok during its peak growth years was for “entertainment” and “distraction.” This is a very different mindset than someone looking for a professional service or a high-ticket item.

I managed a lead-generation campaign for a B2B software company that insisted on being on the platform. Despite our best efforts to match the “vibe” of the app, the leads we generated were often low-quality. The audience demographic trends showed a heavy lean toward Gen Z and younger Millennials who were not the decision-makers for enterprise software.

Platform-Native Ad Placements and User Intent

The “intent” of a user is what they are trying to achieve when they open an app. On TikTok, the intent is passive consumption. This makes “direct-response” marketing—where you ask someone to buy something immediately—very difficult. I found that the platform-native ad placements were excellent for brand awareness but struggled with conversion.

When we look at cross-platform marketing, we have to ask where the budget is most efficient. If my goal was to sell a $500 course, I found that the cost-per-acquisition on TikTok was often three times higher than on other networks, even if the “cost per thousand impressions” (CPM) was lower. You might be reaching more people for less money, but those people aren’t buying.

1. Demographic Target-Matching: This is the process of ensuring the people seeing your ad actually have the money and interest to buy your product.
2. Contextual Targeting: Showing ads based on the type of content a user is currently watching rather than their personal history.

Why Conflicting Algorithm Updates Complicate Budgets

In this section, we explore the lack of transparency in platform updates and how they disrupted agency workflows. We look at the technical challenges of interpreting data when the “rules of the game” change without notice.

One of the most frustrating aspects of managing these portfolios was the frequency of unannounced algorithm shifts. I remember a specific week in early 2022 when every client’s organic reach dropped by 60% overnight. There was no policy update or public statement. We were left to explain to demanding clients why their performance had cratered while our strategy remained the same.

This lack of stability makes it nearly impossible to formulate a real placement blueprint. A blueprint is a long-term plan for where and how you will show your ads. If the platform changes how it weighs “shares” versus “comments” every two weeks, your blueprint becomes obsolete before the ink is dry.

Troubleshooting Metric Discrepancies

Another hurdle was the “attribution” gap. Attribution is how we track which ad led to a sale. Because TikTok is a mobile-only, “walled garden” app, it was notoriously difficult to track users once they left the platform to visit a client’s website. I often saw a 30% to 40% discrepancy between what the TikTok dashboard claimed in sales and what the client’s internal system actually recorded.

• API Integration Shifts: Changes in how the platform’s data “talks” to other software, often breaking tracking links.
• Cookie-less Tracking: The move away from traditional web tracking, which made it harder to prove that a TikTok view led to a purchase on a desktop computer.

Strategic Recommendations for Resource Allocation

This section provides a framework for how agencies can evaluate whether a platform is worth the investment. We focus on a balanced approach to budgeting and the importance of setting realistic expectations for clients.

Based on my longitudinal tracking, I recommend a “60/40” budget split for most growing agencies. This means putting 60% of your resources into a “lead channel” that provides predictable returns and 40% into “secondary support” channels like TikTok. This allows you to experiment with the high-upside potential of short-form video without risking the client’s entire bottom line.

I once worked with an agency that moved 100% of their client’s budget into TikTok because of a single viral month. When the algorithm shifted the following month, the client lost 80% of their revenue and fired the agency. It was a harsh lesson in the dangers of over-leveraging a volatile platform.

A Framework for Asset Customization

To survive the creative demands, agencies must find ways to repurpose content without losing the “native” feel of the platform. You cannot simply take a TV commercial and post it. It has to look like it belongs in the feed.

1. The Hook (0-2 Seconds): You must grab attention immediately or the user will swipe.
2. The Value (2-10 Seconds): Provide entertainment or information quickly.
3. The Call to Action (End): Tell them exactly what to do next, but keep it low-pressure.

Practical Tools for Managing the Chaos

Managing a diversified portfolio requires specialized tools to stay organized. When growth on the platform became difficult, these five tools were essential for my team to maintain some level of sanity and reporting clarity.

1. Vertical Video Editors (CapCut/Adobe Premiere Rush): These allow for quick, mobile-friendly editing that mimics the platform’s native style.
2. Trend Aggregators: Tools that track which songs and challenges are rising in real-time, helping you catch waves before they peak.
3. Cross-Platform Unified Report Cards: Custom dashboards (like those built in Looker Studio) that pull data from multiple sources to show a “true” ROI.
4. Content Scheduling Dashboards: Tools that allow you to batch-upload content, which is vital when you are producing multiple videos per day.
5. Audience Overlay Analysis Tools: Software that helps you see if the people following you on TikTok are the same people buying from your website.

Setup Verification Checklist

Before launching a campaign, I always use this checklist to ensure the foundation is solid. Skipping these steps is a common rookie mistake that leads to wasted budget.

• [ ] Is the TikTok Pixel installed and firing correctly on all checkout pages?
• [ ] Does the creative asset have “safe zone” margins so text isn’t covered by the UI?
• [ ] Is the “Spark Ad” authorization code active for influencer collaborations?
• [ ] Have we set a “bid cap” to prevent the algorithm from overspending on low-quality traffic?
• [ ] Is there a plan for responding to comments within the first four hours of posting?

The Reality of Holistic ROI

In conclusion, the difficulty of growing an agency on this platform stemmed from a clash between traditional business needs and a new, hyper-volatile media format. While the potential for massive reach was real, the lack of predictability, the high cost of content, and the tracking gaps made it a “high-maintenance” channel.

For marketing managers, the key is to treat it as a high-risk, high-reward component of a larger strategy. Never let the “hype” of a viral video distract you from the hard data of conversion rates and customer acquisition costs. By maintaining a data-driven approach and a balanced budget, you can navigate these hurdles without putting your agency’s reputation at risk.

FAQ

Why did my organic reach suddenly drop even though my content stayed the same? This is usually due to an algorithm update or a shift in user “saturation.” If too many people start making similar content, the recommendation engine raises the bar for what it considers “engaging,” meaning your old style may no longer meet the threshold for the “For You” page.

What is a “good” cost-per-click (CPC) on TikTok? While it varies by industry, a healthy CPC for a broad audience is typically between $0.20 and $0.50. If your CPC climbs above $1.00 for a general consumer product, it is usually a sign of creative fatigue or a poor audience-to-product match.

How often should I change my ad creative? In my experience, you should refresh your top-performing ads every 7 to 14 days. Unlike other platforms where ads can last months, the high-consumption nature of short-form video means audiences tire of visuals much faster.

Is follower count a reliable metric for agency success? No. Because of the interest-based algorithm, followers do not guarantee views. I have seen accounts with millions of followers struggle to get 10,000 views on a video. Focus instead on “average watch time” and “conversion rate” as your primary KPIs.

What is the “Safe Zone” in video editing? The Safe Zone refers to the areas of the screen not covered by the app’s interface (like the “Like” button, captions, and profile icon). If your text or call-to-action is in these areas, users won’t be able to read it, which kills your conversion rate.

Can I use the same videos I use on other platforms? Technically yes, but it often fails. TikTok users are highly sensitive to “recycled” content. For the best results, you must use platform-native features like the built-in text styles, trending sounds, and a less-polished, “behind-the-scenes” feel.

Why are my TikTok ads showing a high “View-Through” rate but low sales? This often happens because the platform is great at capturing attention but the “friction” of leaving the app to buy something is high. Users are in an “entertainment” mode, not a “shopping” mode. This is why brand awareness is often a more realistic goal than immediate sales.

What is the difference between an organic post and a Spark Ad? An organic post is a video you upload to your profile for free. A Spark Ad is a paid format that uses an organic post (your own or a creator’s) as the ad creative. Spark Ads generally have a higher engagement rate because they look more “real” to the user.

How do I justify the high production costs to a client? Focus on the “testing” aspect. Explain that by producing more low-cost, raw videos, you are finding the “winners” faster. One successful video can often pay for the production of fifty “failed” ones if the ROI is tracked correctly.

What is “Shadowbanning” and is it real? While the platform doesn’t use the term “shadowban,” they do have “visibility filters” for content that violates community guidelines or is deemed low-quality. If your views drop to zero suddenly, check if your content is being flagged for “unoriginality” or sensitive topics.

Does the time of day I post matter? Less than it used to. Because the algorithm shows content based on interest rather than chronology, a video can “go viral” three days after you post it. However, posting when your specific target audience is most active can help provide the initial engagement “spark” the algorithm needs.

(This article was written by one of our staff writers, Jonathan Mercer. Visit our Meet the Team page to learn more about the author and their expertise.)