Three’s Company Media Group Co., Ltd. (referred to as “Three’s Company” or “the company”) announced on March 7, 2026, that it intends to change its Chinese name to “Three’s Company Future Technology Group Co., Ltd.” and its English name to “Three’s Company Future Technology Group Co., Ltd.” The company also plans to reclassify its industry from “Business Services” (L72) to “Internet Data Services” (I6450).

Three’s Company explained that the name change aligns with the implementation of its current strategy and lays the groundwork for future development, matching the company’s core business and strategic positioning. It emphasized that “there is no intention to influence the company’s stock price or mislead investors through the name change.”

Since 2023, regarded as the “first year of generative AI applications,” Three’s Company has increasingly referenced “AI” in its annual reports. However, research and development expenses have declined over the past two reporting periods, falling by 1.54% and 7.54%, respectively, with the rate of decrease accelerating noticeably.

Meanwhile, the embrace of AI does not appear to have positively impacted the company’s performance. From 2023 through the first three quarters of 2025, both revenue and net profit attributable to shareholders have continued to decline. Net profit margin and return on equity have hit record lows since the company’s listing, diverging from the high growth seen in the industry.

A review of financial reports from several digital marketing companies reveals that their so-called AI marketing products are often merely interfaces built on top of existing large language models, relying heavily on the same few major model providers. This approach fails to establish a solid technological moat. If the underlying models are upgraded or their services become more widely accessible, these companies’ businesses could face significant disruption.

Although Three’s Company has significantly increased its mentions of AI in financial reports, its operational performance has begun to deteriorate.

Three’s Company is a comprehensive advertising and media enterprise specializing in integrated marketing services. It provides clients with digital marketing, scenario-based event services, and campus media marketing, enabling multi-scenario, wide-coverage, and high-efficiency dissemination of marketing information to meet cross-media, multi-channel brand exposure and product promotion needs nationwide.

In recent years, the explosive application of AIGC (Artificial Intelligence Generated Content) has profoundly reshaped the production and delivery models within the advertising industry. Breakthroughs in deep learning and large model technology have not only enabled intelligent generation of advertising creatives but have also pushed the precision of personalized ad delivery to unprecedented levels.

In 2023, marked as the first year of generative AI application, Three’s Company notably increased its references to “AI” in periodic reports.

The 2023 Annual Report mentioned “AI” 14 times, compared to just 3 times in 2022. During the reporting period, Three’s Company signed a strategic cooperation agreement with iFlytek to enhance smart marketing capabilities through AI empowerment and jointly develop multi-modal intelligent marketing tools based on next-generation AI technology. By the end of the year, it officially launched an AI tool for marketing called “One” AI.

The 2024 Annual Report mentioned “AI” 18 times, an increase of 4 mentions year-over-year. During this period, the company upgraded “One” AI from version 1.0 to 2.0 and launched an AI event prediction platform by year-end.

The 2025 Interim Report mentioned “AI” 40 times, approximately double the number from the full year 2024. The report indicated that Three’s Company continued to iterate, upgrade, and optimize its multi-modal AI products for marketing, successfully building the “One” AI full-link marketing intelligence system. A self-developed sports event intelligence robot had also completed its prototype.

However, the reality has fallen short of ambitions. As Three’s Company vigorously embraced the AI wave, its operational performance began to decline.

In 2023, Three’s Company reported revenue of 5.284 billion yuan, a year-on-year decrease of 6.53%, and net profit attributable to shareholders of 528 million yuan, down 28.36% year-on-year. This marked the first decline in performance since its IPO, breaking a 10-year streak of continuous growth records dating back to 2013.

In 2024, the company’s performance declined further. Revenue fell to 4.208 billion yuan, with the rate of decrease widening to 20.36%. Net profit attributable to shareholders dropped to 123 million yuan, plunging 76.65% year-on-year. The net profit margin was only 2.92%, lower than even 2012 levels.

In the first three quarters of 2025, Three’s Company failed to reverse the downward trend. Revenue decreased by 16.72% year-on-year to 2.569 billion yuan, while net profit attributable to shareholders fell 20.48% year-on-year to 144 million yuan. The company’s performance forecast projects full-year 2025 net profit attributable to shareholders between 181 million yuan and 209 million yuan, representing a year-on-year increase of 46.79% to 69.50%. However, adjusted net profit attributable to shareholders is forecasted between 189 million yuan and 247 million yuan, indicating a potential change ranging from a decrease of 19.63% to an increase of 5.04%.

Regarding the anticipated profit growth in 2025, Three’s Company stated that while ensuring high-quality development, it remains committed to leveraging AI innovation to drive performance growth. It claims that applying AI technology has improved staff efficiency, leading to cost reduction and efficiency gains reflected in its financial results.

But is this truly the case? Three’s Company first mentioned AI in its 2022 Annual Report. That year, its gross profit margin was 16.68%, a significant decrease of 4.47 percentage points year-on-year, while its net profit margin was 13.02%, down 1.12 percentage points.

In 2023 and 2024, although the company’s gross profit margin slightly recovered, it fluctuated only between 17% and 18%, failing to return to above 20%. The net profit margin plummeted to 2.93%. In 2024, the company’s ROE (diluted) dropped sharply to 4.59%, falling into single digits for the first time with a decrease of 13.52 percentage points.

In the first three quarters of 2025, Three’s Company’s gross profit margin and net profit margin were 17.18% and 5.63%, respectively, down 0.34 percentage points and 0.27 percentage points year-on-year. Profitability has not substantially recovered. ROE declined further, dropping 1.12 percentage points year-on-year to 5.39%.

According to Wind data, from 2023 to the first three quarters of 2025, total revenue for the advertising and marketing industry (Shenwan classification) increased year-on-year by 7.17%, 3.87%, and 7.19%, respectively. During the same periods, BlueFocus saw revenue growth of 43.44%, 15.55%, and 12.49%; Focus Media reported growth of 26.30%, 3.01%, and 3.73%; and Gravity Media achieved growth of 9.34%, 32.73%, and 45.78%.

Compared to its peers, the results Three’s Company has achieved in the AI field do not appear to have translated into meaningful contributions to its performance.

Despite proclaiming a full embrace of the AI wave, Three’s Company’s sales expenses remain significantly higher than its R&D expenses. Furthermore, R&D expenses have declined over the last two reporting periods.

From 2016 to 2022, Three’s Company’s R&D expenses increased from 796,300 yuan to 42.035 million yuan. Although this represents a 41-fold increase, the investment remains relatively low compared to revenues in the billions, with an R&D expense ratio of just 0.74%. In 2023, R&D expenses rose to 53.8821 million yuan. Coupled with declining revenue, the R&D expense ratio increased to 1.02%, surpassing 1% for the first time.

However, in 2024 and the first three quarters of 2025, while mentions of AI increased substantially, R&D expenses continued to fall, declining by 1.54% and 7.54% respectively, with the rate of decrease accelerating. Nevertheless, because revenue declined at a faster rate, the R&D expense ratio actually increased.

Currently, the digital marketing industry exhibits a distinct pyramid structure. At the apex are giant technology platforms like Tencent, ByteDance, Google, and Meta, which control core algorithms, data, and AI technologies, and invest heavily in R&D. For example, Meta’s R&D expenses from 2023 to 2025 were $38.483 billion, $43.873 billion, and $57.372 billion respectively, with an R&D expense ratio approaching 30%.

The vast majority of listed digital marketing companies are essentially “service providers” or “agents” that depend on these major platforms. They directly use mature tools provided by the platforms (such as ByteDance’s UBMax or Tencent’s Advertising 3.0 system) to serve clients. Consequently, they neither need nor have the capacity to conduct large-scale underlying technology R&D themselves.

To escape homogenized competition, digital marketing companies are increasing their focus on AI and launching self-developed AI marketing products.

BlueFocus released its industry model, BlueAI, which has now achieved scaled implementation in overseas and domestic marketing scenarios, incubating dozens of vertical agents for core segments like integrated marketing and influencer marketing. Focus Media has completed the deployment of its marketing-specific large model, which entered the testing phase as of January this year. Its internally developed AI agent, “Zhong Xiao Zhi,” is already in application. Gravity Media introduced its own marketing and e-commerce industry model and launched three intelligent agents: Perception Agent, Decision Agent, and Content Agent. It also released three AIGC product applications based on actual business scenarios: “Creative Assistant,” “Video Assistant,” and “Universal Assistant.”

Recently, a surge in the A-share AI application concept fueled significant gains in related digital marketing stocks. From November last year to January this year, Three’s Company’s stock price rose over 60%. Between January 5 and 13, Gravity Media recorded six limit-up trading sessions in seven days, with its stock price surging over 80%. From January 9 to 14, Tianlong Group saw three limit-up sessions in four days, with its stock price increasing over 90%.

However, industry insiders emphasize that the underlying technology for these AI products almost universally relies on the same few major model providers (such as Deepseek, Doubao, Wenyan Yixin, and Zhipu). Their core advantage lies mainly in their deep understanding and adaptation to vertical business scenarios, rather than originality in the underlying models, which does not constitute a solid technological moat.

Many digital marketing companies’ AI products are essentially “wrappers” around underlying large models. If the capabilities of these underlying models are upgraded or their services become universally accessible, these companies’ businesses could be vulnerable. The struggles of the US AI marketing company Jasper AI serve as a cautionary tale.

In 2021, Jasper AI, leveraging GPT-3, launched a marketing tool via API. After iterations, the product was offered as a SaaS-like service. By 2022, Jasper AI’s revenue reached $80 million.

However, the success was short-lived. Following the launch of ChatGPT in 2022, traffic to Jasper’s website declined sharply, and revenue dropped significantly. Other companies like the no-code AI marketing platform Mutiny and the speech recognition AI firm Deepgram also subsequently encountered developmental difficulties.

Factors such as ChatGPT’s direct market entry, competition from tech giants like Google and Microsoft, and pressure from the open-source ecosystem have weighed heavily on companies like Jasper. Beyond these external factors, a fundamental reason is that these companies failed to build a moat around their large model capabilities.

A technology investor from a top-tier investment institution commented, “Jasper’s business did not create or meet a genuine need. When large model capabilities became stronger, their business was naturally superseded. Open AI’s core strength is its large model capability; companies like Jasper merely followed the trend by creating wrappers without creating real value.”

For digital marketing companies like Three’s Company, the key challenge in the next phase will be whether they can leverage the current AI红利 to accumulate truly valuable, difficult-to-replicate assets. This will determine their ability to break free from homogenized competition and build a sustainable competitive moat.