How To Do Better Influencer Marketing

Agung Dwi Sandi is the Founder and CEO of rankpillar Group, an expert marketing and leading digital marketing agency in Southeast Asia.

gettyInfluencer marketing has stopped being an experiment. According to the World Federation of Advertisers’ 2025 research, 99% of member brands work with influencers, 66% do so frequently and global creator partnership spend is expected to pass $39 billion this year. The discipline has professionalized quickly on the compliance side: 70% of brands say they always sign contracts, and 66% have a written influencer policy.

What has not professionalized at the same rate is what brands actually ask for.

I run a digital marketing agency in Southeast Asia. Across the region, a large paid-amplification economy has grown up. In Indonesia, it is commonly called the buzzer industry: loosely organized networks of social accounts, assembled by coordinators, that will post on a brand’s behalf for a fee priced per account, per post, per day.

Comparable structures exist in most high-engagement social markets. Indonesia’s is simply unusually visible and unusually cheap. Enforcement is now almost entirely automated: KOL.ID’s read of transparency data puts TikTok’s proactive removal rate in Indonesia at 99.3%.

The briefing conversations I have with brands buying this kind of amplification follow one script. How many accounts can you mobilize? How many posts a day? Can we be trending by Friday? What almost never comes first is the only question that decides whether any of it works: What should a specific person believe or do differently after they see this?

That gap is structural rather than ethical. Read what this part of the market charges for, and you have read its self-definition: accounts, posts, days and impressions. Nothing on that price list is an outcome.

Buyers are not innocent either. In the same WFA research, reach (94%) and engagement (89%) dominate success measurement, while only 23% of brands use conversions or sales. Marketing Dive reported this month that about two-thirds of marketers surveyed by the Association of National Advertisers call measurement the hardest part of influencer work.

When a buyer cannot measure an outcome, they buy what they can count. Volume is countable, so volume becomes the product.

There are four key changes that I believe will separate brands that buy influence from brands that buy noise.

A brief is not a document; it is a set of decisions: Who are we talking to, what do they currently believe, what do we need them to believe instead, what are we asking them to do and what evidence would we accept to show that it worked?

Hand a vendor 200 posts to place, and you have given them a logistics problem. Logistics firms are good at logistics. What they cannot do is tell you that your positioning is the thing that is broken. If the first thing you receive back is a rate card, you were never buying marketing.

The WFA’s own recommendation is to look past follower counts and look toward audience alignment and genuine product affinity.

In our creator work, candidates pass three gates before price is discussed: whether their audience actually sits in the geography the client sells in, whether their recent short-form content earns real view-through rather than borrowed spikes and whether the content would survive being placed next to the brand’s own. We source candidates through application-based creator hubs, where creators opt in to a campaign rather than being bulk-booked by a coordinator, which makes consent and fit inputs to the brief instead of afterthoughts.

Matching is slower and far harder to sell than mobilizing a hundred accounts. It is also the only part of the process that compounds.

The ANA’s guidance is to set an end goal, whether it’s sales, conversion or awareness, before launch. Go one step further and define the disproof. A credible answer involves a holdout group, a before-and-after measure or, at minimum, a number agreed upon in writing that nobody can reinterpret afterward.

Ask a prospective partner what they would accept as evidence the campaign failed. The quality of that answer tells you more than any case study.

The ANA recommends assigning a single internal group to manage creator relationships, and the WFA finds marketing departments leading this work in 54% of organizations.

The question to settle internally is simple: Who writes the brief? “Send us the brief, and we will execute” is a supplier’s answer. That may be exactly what you need, but do not then expect the result to be a strategy. Ask a partner, too, what they would remove from your plan. A labor vendor will add scope because scope is revenue. A partner will subtract whatever does not serve the objective.

Generative tools have collapsed the marginal cost of producing and distributing content with streamlined drafting, variation, translation, scheduling and just sheer volume. A business whose value proposition is the supply of execution labor is being repriced in real time, and a brand whose strategy is the purchase of execution labor is buying an input its competitors can now acquire for almost nothing.

What does not commoditize is judgment about an audience, a point of view worth amplifying and the instrumentation to know whether either worked. In the campaign data I see, the distance between paid volume and genuine attention has widened over the past two years, which makes a warehouse of accounts a depreciating asset rather than an unfair advantage.

The amplification economy did not grow because marketers in my market are careless. It grew because it answers, cheaply and instantly, a question that is easy to ask: How much noise can I buy?

The harder question has no rate card: What is the one thing people need to understand, and how will we know they understood it? Write that down before you brief anybody. If a vendor cannot argue with your brief, they cannot improve your marketing. They can only make it louder.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?