100K+ Users in Just 15 Days After Launch: How CodeFlying Turned Early Momentum into Scalable Growth

For most teams building coding related products, YouTube is usually the first platform that comes to mind. The content format is longer, the explanations are more complete, and the user journey tends to follow a more natural pattern of understanding first, then converting.

But CodeFlying came to a different conclusion.

When talking about influencer performance, Zhihao Xu, who leads marketing, shared that overall, Instagram delivers better conversion and higher quality users than other platforms. When they broke it down further, their current ranking was clear: Instagram first, YouTube second, TikTok third.

This was not a generic industry assumption. It was a conclusion CodeFlying arrived at gradually through actual campaign execution and repeated post campaign analysis.

Meet CodeFlying

CodeFlying is a no code app generation tool. Users describe what they want in natural language, and AI generates a mobile app that is ready to publish. There is no code interface and no technical barrier. In their own words, even a 60 year old grandmother or a young child could use it.

The product was originally launched in China under the name 码上飞, targeting the domestic market. At that time, there were still very few direct competitors in China, and the team used that window to grow to 700,000 users, with a stable paid conversion rate of 6% to 7%.

But they knew very clearly that this window would not stay open forever. As Lovable entered China and the term vibe coding became increasingly popular, competition in the market intensified quickly. That was when they shifted their focus overseas.

Their decision to make the product mobile first also came from what the team had learned in China. As Zhihao put it, the mini program ecosystem had already proven that mobile products are naturally more shareable than web based ones, and they can be used anytime, anywhere.

In December 2025, CodeFlying officially launched. In just half a month, it surpassed 100,000 users and ranked second on China’s AIGC Rank list for fastest growing AI websites. Zhihao and one other operations teammate moved quickly, testing and refining growth paths in real time.

New products have an early momentum window, and influencer marketing is the fastest way to amplify it

After launch, the team focused on a few key levers: influencer marketing, Google and Meta ads, and some media PR. Together, those three moves gave them a strong start.

For them, paid ads serve two main purposes. The first is capture. That mainly means search ads, including competitor keyword bidding and branded keyword defense. The second is amplification. They take influencer content and run it as paid media to extend its reach and compound the value of the content.

As Zhihao explained, ads and influencer content are not competing with each other. They work together. Influencers create the content, and then the team uses paid media to push it further. Or, once influencer content starts building awareness, search ads are there to capture demand more accurately.

But the real growth engine was influencer marketing.

As Zhihao put it, every new product has its own early momentum window, and influencer marketing is the fastest way to amplify that momentum.

For a product in the vibe coding category, the core value is inherently something users need to see in action. An influencer says one sentence to the camera, opens CodeFlying, and a few minutes later a real app appears. That visual transformation is the product story. Traditional ads simply cannot carry that same level of information density. Influencer content can.

That is also why their core budget and internal priority now lean more heavily toward influencer marketing. Compared with other channels, the ROI is clearly stronger.

Before working with AhaCreator, the team ran an in house plus agency model in parallel.

Each approach had its own pros and cons, but both shared one unresolved problem: unreliable data.

Zhihao said that everyone uses some kind of third party tools, but the influencer data and audience insights in those tools are often inaccurate, and the communication process is also cumbersome. When so many uncertainties stack on top of each other, it becomes very hard to tell whether each dollar is being spent in the right place.

The other issue was speed.

When they ran a campaign themselves, the full process from sourcing influencers and outreach, to negotiation, scheduling, and final delivery usually took at least three to four weeks. That meant they had no way to move quickly or learn quickly.

A small team, but still running 70 to 80 influencer collaborations per month

Zhihao first learned about AhaCreator through an article on WeChat. Once the team actually started using it, the two biggest changes were immediately clear: the cycle became much shorter, and execution pressure dropped significantly.

First, the timeline.

Zhihao gave a very clear range. Compared with doing everything in house or working through an agency, the overall campaign cycle was shortened by roughly 30% to 40%. In his example, a campaign that used to take at least three to four weeks from start to finish could now move from influencer kickoff to content delivery in as little as one to one and a half weeks under a faster execution rhythm.

Then there was execution.

He described the daily workflow very simply: they basically check once a day, handle anything that needs action, and close it if there is nothing new. It does not take much time. Even if they do not check for a few days, the platform will proactively send email reminders when something needs attention, so it is easy to stay on top of things.

In other words, the team only needs to step in at key decision points.

Under their previous workflow, they were chasing email threads almost every day, confirming delivery status, pushing timelines, and manually tracking progress across individual influencers.

He also called out one specific detail: content usage rights and ad authorization codes no longer needed to be negotiated separately with influencers, which he felt was a major advantage of AhaCreator. Under a normal agency model, those ad authorization codes often require extra payment. On the platform, there is no need for separate negotiation. Once the influencer content is published, the codes are available right away.

Because execution costs dropped so much, a five person team is now able to consistently maintain 70 to 80 influencer collaborations per month.

And once that level of volume is possible, post campaign analysis starts to become truly meaningful.

The practical judgments CodeFlying has gradually developed through influencer marketing

1. Influencer selection: filter out fake data first, then evaluate ROI potential

When it comes to selecting influencers, they have learned the hard way what inflated metrics look like. Some influencers appear to have strong numbers, but do not drive any real conversion at all. The overseas market is full of noise. AhaCreator helps filter out influencers with fake data or suspicious traffic patterns at the selection stage.

That gives the team a cleaner pool to work from. After that initial filter, their main criterion becomes simple: ROI potential.

For example, if an influencer charges $1,000 but is projected to deliver a CPC below $0.50, that is still worth doing. Some influencers may not have huge followings, but if they consistently produce breakout content, those are exactly the kinds of partners the team prefers.

2. Reinvesment decisions: CPM is the first filter, conversion is the second

Their logic for repeat investment is highly structured. First they look at CPM. If CPM is too high, they eliminate the influencer immediately, because that usually means the content itself lacks distribution power. If CPM passes the first check, they then look at conversion. Influencers with strong conversion get reinvestment. Those with weak conversion are dropped.

There is one exception. If CPM is high but conversion is exceptionally strong, they will still consider continuing the partnership. As Zhihao explained, if the lower funnel results are strong enough, they are still willing to keep the influencer.

3. Campaign structure: always on testing plus feature launch campaigns in parallel

Their campaign rhythm now runs on two parallel layers.

The first layer is always on testing. These campaigns are split by region and platform. For example, they may run one campaign specifically for North America and another specifically for Europe, with data tracked separately and conclusions built separately rather than blended together. The purpose at this layer is to continuously improve their judgment around influencer selection and content direction.

The second layer is campaign work tied to product iteration. For example, they have an important new feature launching in April, so they are building a campaign in advance around that feature, identifying influencers who are a strong fit for that product story, and concentrating publication around the launch window so influencer content aligns with product timing.

4. Platform selection: let data override assumptions, and Instagram comes out on top

For influencer marketing around tools products, many teams instinctively bet on YouTube first because the format allows for more detailed explanation and a more rational decision making process. But CodeFlying’s actual results told a different story: Instagram short form video ranked first, YouTube second, and TikTok third.

After seeing the data, they developed their own interpretation. CodeFlying’s core users are people trying to build practical tools that can monetize directly. That audience is not any less concentrated on Instagram than it is on YouTube. At the same time, distribution on Instagram depends more heavily on follower quality and trust, which tends to produce more precise users than TikTok’s algorithm driven traffic.

5. Content: the hook matters more than production quality

On the content side, one insight became especially clear after repeated analysis: the hook matters far more than production quality.

They ran one video that performed extremely well across every platform, with five to six million views on YouTube and several million on Instagram. The production quality was not especially polished, and the editing was not particularly sophisticated either. The only thing that really made it win was the opening.

At the start of the video, the influencer said something with strong contrast value, essentially: stop paying people to build software for you. Watch this. I just made one in a few minutes.

When Zhihao analyzed that video, his conclusion was simple: a strong hook is already enough to pull people in and make them remember the product. Compared with that, the middle section of the tutorial matters much less.

That insight has now been built directly into their brief strategy.

They do not require highly polished production. They do not tightly control the influencer’s content structure either. What they care about is whether the influencer can come up with a strong hook and whether the video makes the viewer want to try the product.

When a hook performs well, they continue refining it and making it available for other influencers to reference and adapt.

Their review standard is also very simple: from the user’s point of view, can someone stay through the first five seconds, and after watching the video, do they feel the urge to try it themselves?

They do not overanalyze structure or nitpick details. That is the one question they keep coming back to.

Nearly 2M impressions, with CPC as low as $0.95

According to AhaCreator platform data, this campaign generated nearly 2 million impressions, with CPC as low as $0.95.

One influencer alone delivered more than 1,300 clicks on a collaboration fee of just $110, bringing CPC down to only $0.09.

Those numbers were not luck.

They were the result of the team stacking three judgments on top of each other within a limited budget: choosing the right influencers, finding the right content angle, and prioritizing the right platforms.

As those decisions continue to accumulate and improve across campaign after campaign, influencer marketing has evolved for CodeFlying from something they were trying experimentally into a growth engine they can keep optimizing and scaling over time.

AhaCreator ToolkitYour account portal for all AhaCreator tools