Working with creators

This article is part of a series of articles from the WARC Guide to working with creators. Read more

It was first published as part of WARC China’s Spotlight on the influence and potential of KOL/KOC.

Why it matters

In an economy where brands are working ever closer with large numbers of creators, virtual influencers present a less-risky alternative, but marketers must still be careful when using them.


  • Virtual KOLs have become the new favorites of marketers.
  • Virtual KOLs come with their own advantages and disadvantages in terms of target audience, price and form of interaction.
  • Virtual KOLs are less risky than real-life KOLs or celebrities, but are not completely risk free.
  • Real-life and virtual KOLs both need reasonable measurement tools to reduce the risk of partnership.
  • The measurement and evaluation methods of virtual KOLs are similar to that of real-life KOLs; overall, it depends on the purpose of the placement.

Virtual KOLs have become the new favorites of marketers, but advantages vary

Although many marketers are eager to try virtual KOL (Key Opinion Leader) marketing, or have already tried it on a small scale, when it comes to planning, virtual influencers need to be compared with potential real-life KOLs to determine if they are the right fit for overall marketing goals and appropriate for the specific format of placement.

From a category perspective, real-life KOLs in each vertical field are proving themselves in relevance and performance, with a sufficient number of brands available that they can partner with. The target audiences of virtual KOLs are predominantly Gen-Z, who are ready to engage with virtual avatars for their trendy and fashionable nature. However, the relevant categories they can work with are limited to Beauty, Food, Automotive and 3C (computer, consumer electronics and communication products).

Graphic representation of self-published media by channel and category

KOL placements by marketers are distributed across all categories, including Beauty, Video, 3C, Mother & Child, Automotive and Finance.
Source: R3 China KOL Market Analysis Report.

In terms of interaction, virtual KOLs are not restricted by physical or geographic limitations and can appear in various types of marketing scenarios. Virtual characters can also have similar backgrounds, looks, ‘experience’ and fan operation teams to that of a real person. Although they need to be coordinated and communicated via their agents/companies, the time and effort required to activate virtual influencers are generally less than that of real-life KOLs or celebrities.

Types of virtual KOLs and audiences

Source: R3 Content Radar 2020 Summary Report; The Rise of Virtual Influencers: An R3 Report.

However, it should be noted that while virtual KOLs are often viewed as a safer alternative to celebrities or real-life KOLs, they are not immune to scandals. In May 2022, A-SOUL member Jiale announced the termination of her live broadcast. The fall-out had a negative impact on the group and the contracted company, which led to widespread discussion about the employment relationship between the person behind the avatar and the company she was contracted with.

It is important for marketers to keep in mind that virtual KOLs are ultimately controlled by real people. Whether it is an actor or an agency, risks can exist in the form of personal disputes, labor relations, and capital operations. Regardless of the form (virtual or real-life) an influencer takes, brands need to establish how they measure the use and effectiveness of their KOLs.

Anime characters

Source: A-SOUL.

Improving KOL return on investment

The cost of virtual KOLs can be comparable to some top-tier KOLs and even some celebrities. As a result, brands need to carefully consider how they facilitate the decision of whether to invest in influencers and at what price. Marketers need to measure the performance of their influencers at different stages of KOL placement. Establishing a planning logic for the end-to-end lifecycle of KOL engagement and ensuring that the right measurement tools and reporting is in place are the only ways to avoid risks.

Graphic showing strategy workflow

Source: 2022–2023 China KOL/Short Video Marketing and Management: An R3 Report.

To measure KOL performance effectively, marketers need to be aware of the management problems that are often encountered in the KOL placement process. Some of these issues are:

  • Irrational buying model, buying agencies that do not match the demand, no reasonable standard for buying service fee: Buying efficiency, management efficiency and price competitiveness are not enough to support the current stage of marketing business needs.
  • No regular and reasonable benefit verification system: No way to verify whether the benefits are implemented on time, in sufficient quantity, and in a qualified manner.
  • No database or low standardization of purchasing data: No scientific and reasonable comparison, analysis and external benchmarking.
  • No reasonable intermediary restraint and control system and terms: Loss of efficiency and marketing expenses cannot be traced.

These problems can make it difficult to effectively manage KOL marketing and ultimately achieve accountability for marketing spend. By optimizing the management of the above processes, marketers can clarify the selection of KOLs and track actual results.

Using real-life KOL methods to measure virtual KOL performance

The methods associated with measuring real-life KOL performance is relatively mature in China. Similar measurement approaches can be directly applied to virtual KOLs, with the exception of two scenarios.

Scenario one: The marketer builds a virtual image of a real person

Compared to using real people, using athletes’ identities in the virtual world can attract a large audience. Their public persona and existing number of fans and followers can deliver greater brand awareness and visibility for marketers. The use of virtual influencers also offers the flexibility to apply innovative ideas without the constraints of the real world (i.e., ability to be in attendance at one place at one time).

Virtual KOL ads

Source: Migu Video, China Mobile.

China Mobile’s Migu Video and the Olympic skier Eileen Gu collaborated on ‘MeetGu’, an avatar created so that Gu could participate in various promotions during the Winter Olympics in China; which she was also competing in. Digital images were created of Gu for its campaign promotion. Users could interact with MeetGu in an immersive and real-time manner. In addition, MeetGu served as a commentator for the Olympic skiing event, a reporter for the event broadcast and a host for the e-commerce live stream. To break the third dimension and increase awareness, China Mobile also created naked-eye 3D ads that were advertised on large outdoor advertising screens in many central business districts in China.

In this sense, the application of virtual image is more of creative communication and IP extension, without independently establishing social accounts for the virtual image or operating to accumulate fans. The characteristics of this type of engagement are more similar to celebrity endorsements or an extension of a creative partnership, and cannot be completely evaluated as KOL marketing.

Scenario two: Virtual KOLs that have established their accounts on social platforms and have been operating for a considerable time

Virtual KOLs can be measured by communication indicators that are aligned with the characteristics of individual social media platforms. Specifically, since each platform has different positioning, target audience and distribution mechanisms (listed in the table below), brands need to measure the ability of KOLs with consideration given to different interactive dimensions per platform. At the same time, rating mechanisms should be introduced to better judge the effect of KOL placement when measuring the effect of the campaign.

Table showing channel attributes

Source: 2022–2023 China KOL/Short Video Marketing and Management: An R3 Report.

The combination of real and virtual KOL activities is essentially a brand marketing campaign, so it should be evaluated on the purpose of the brand marketing campaign and its desired effect. Just like an O2O marketing campaign, it depends on whether the brand wants to attract new traffic from online to offline, or whether offline feeds online for the purpose of increasing fans, building community or re-purchasing.

Different campaign communication goals require different measurement indicators:

  • If the focus is brand building, marketers should pay attention to brand health, search volume changes, brand keyword mention changes.
  • If the focus is seeding, marketers should pay attention to communication (exposure/play), interaction volume/rate, sharing, public opinion (positive/negative).
  • If the focus is conversion, marketers should focus on viewing, in-store headcounts, purchase, return.

Although the entities of real-life and virtual KOLs are very different, in the end the brand is working with an extension based on a marketing purpose. And furthermore, the evaluation of marketing performance is also based on its purpose. Most of the time, it is not because of the KOL entity or the digital body that the evaluation criteria of the marketing campaign is affected and changed; or, in this changing era, more often than not, marketers are also seeking common ground in many marketing tools to achieve their objectives.

Landing on the evaluation of each cooperation, especially KOL positioning and performance, on the basis of strategic classification, marketers gradually adopt a more strategic portfolio KOL placement model to generate different brand influence, product seeding power and commercial conversion power at different nodes. In this model, virtual KOLs are similar to the positioning of top-tier to mid-tier KOLs.

Graphic showing relationship between KOLs and categories

Source: China KOL Market Analysis: An R3 Report.

KOL measurement solution from short-term to long-term effect link

Based on R3’s long-term focus on the marketing landscape and its experience in serving marketers, readers may find the following two case studies a glimpse of solutions that are applicable to their current situation:

Case one

A beauty brand was going through a rapid expansion in its KOL buying business. The company originally possessed a scattered KOL marketing strategy, but then shifted to integrate its KOL marketing with one agency as part of a long-term strategy. Based on this clear purpose, the beauty brand underwent a customized pitch management process to find a partner with full KOL chain service capability and a competitive price advantage in at least one specific platform.

In addition, the beauty brand conducted a comprehensive diagnosis of its Group’s existing KOL service control system. This identified problematic issues and areas for optimization in internal and external workflows. It also encouraged greater collaboration between marketing and procurement compliance, helping the Group’s brands and procurement be more transparent about the partnership process. Agency performance evaluation and cooperation efficiency was also significantly improved.

Case two

A multinational luxury group has seen its KOL marketing spend increase in parallel with their growth in China and wanted to evaluate the effectiveness of its KOL investment in terms of both actual performance and deliverables. When it comes to strategy and media planning and buying, the Group works with many agencies and a variety of buying platforms. This has, however, led to problems in areas like the opacity of pricing and discrepancies between deliverables and those stated in buying contracts.

After looking through the group’s KOL spending, we’ve decided to benchmark by product category, agency and social platforms. For KOL placement platforms, most of the social platforms evaluated are measured for effectiveness through CPE metrics, although their key head placement platforms, such as Douyin, Weibo, Bilibili and Red, on which cross-industry brands have varying performance, are benchmarked through internal databases.

Marketers should focus on metrics that are best suited to each platform and industry. For example, Red posts have a higher number of ‘likes’ and ratings. Weibo and Bilibili have the lowest likes rating ratio. When it comes to KOL placements in Beauty and Personal Care categories, CPEs tend to be relatively high. In terms of auditing benefit fulfillment against what is stipulated in KOL contracts, marketers need to be clear of the form of their bought placement, such as the sequencing of WeChat articles and the length of exposure in video content. It is also important to note that when calculating the value of content, that depending on platform, content might be removed and hidden by a KOL after a period.


  • Virtual KOLs have become the new favorites of marketers.
  • Virtual KOLs come with their own advantages and disadvantages in terms of target audience, price, and form of interaction.
  • Virtual KOLs are less risky than real-life KOLs or celebrities, but are not completely risk free.
  • Real-life and virtual KOLs both need reasonable measurement tools to reduce the risk of partnership.
  • The measurement and evaluation methods of virtual KOLs are similar to that of real-life KOLs; overall, it depends on the purpose of the placement.
Specific measurement optimization methods include:
  • Agency cooperation process and procurement transparency.
  • Each brand’s working methods, processes and management tools.
  • Historical buying against market prices and diagnostics done according to platform, marketing purpose and specific indicators.
  • Continuous monitoring of placement through third-party platforms or agencies.
  • Industry best practice learning.

About the author

Read more articles from the WARC Guide to working with creators.

How digital commerce adoption influences the way brands partner with creators in Asia
Wing Lee, Korakan Yamasattham and Shodai Fujita
AnyMind Group

In the battle for attention, content is everything
Jeremy Hollow
Listen + Learn Research