Gen AI will have a profound impact on agency business models

There have been many transitions in the ad industry over the last 40 years. Those old enough will remember the transition into ‘desktop publishing’ (DTP) where manual layouts turned into blueprints got replaced by designing them on a computer. Then came the Internet and with it SEO, ecommerce, social media, programmatic, CRM, mobile, influencer and the like. Every time agencies had to adapt, in technology, skillset, process, organisation and culture.

 

Today, we are fast approaching the next transition: the impact of gen AI. And while the industry runs heated discussions about the impact of gen AI on creativity (as usual), this transition will have a much more profound impact on the foundations: the agency business model.

 

Commercial creativity
 

Let’s cover creativity first. While it’s true that we are a creative industry, there is the delusion that creativity is everything. That is not correct. Creativity in business is a means to achieve commercial goals. Our clients buy ‘commercial creativity’ and for the ad industry, the ‘commercial’ part has been falling short for many years. We celebrate the ‘drill’ at Cannes with our creative work while our clients buy the ‘hole’ or the impact on business.

 

This ‘misunderstanding’ of the value agencies provide to the business now haunts back the industry. The focus on selling creativity has become less and less meaningful to clients. They (have to) look for results.

 

If in doubt, look at the successes of consultancies such as Accenture Song or Deloitte Digital, or of Meta, Google, and the like to cut out agencies and deal directly with clients. Or look at the ever-shrinking share agencies capture from ever-growing marketing budgets.

 

 

Agencies’ current business model

Now, why is gen AI a business-model changer? Agencies have two kinds of projects:

Most agencies love big-idea work as it makes agency staff proud and offers opportunity to promote the agency. But, BIW is rarely profitable anymore: it’s labour-intense, requires senior people to be involved, and many rounds of client interactions. BAU on the other hand is fairly predictable: templates can be used, junior staff can generate the work and it requires little senior supervision.

 

Most agencies run a mixed-model calculation on these two parts of the work, so the less profitable BIW is balanced off by the very profitable BAU.

 

The gamechanger
 

But  gen AI challenges this model. That is, because BAU work can be templated, generated by junior staff, without much supervision... meaning, gen AI can take care of BAU work. Not for the agency, but directly for the client. And clients are realising that.

 

In fact, clients will have to. The time and cost advantages of AI-generated BAU is just too huge to be ignored. Instead of hundreds of dollars and days of work, it’s tens of dollars and minutes of work. Procurement and client’s own CEO (who likely uses ChatGPT for his/her own email messages) will ask the marketing team to utilise the technology. So even if your client doesn’t ‘believe’ in gen AI, they are likely to be forced to explore it.

 

Financial necessities

So, what’s the consequence to the business model? Let’s make a calculation:

 

Assume an agency currently gets paid 30% for big-idea work and 70% for business-as-usual work (and assume BIW generates -20% loss and BAU 20% profit). Such setup generates about 8% profit *(30 x (-0.2) + 70 x 0.2 = 8)

 

Let’s assume the client decides to take half of the BAU work (=35%) in-house. As an effect

So, apart from losing 35% of the original budget, the agency also makes only 1/5th of the profit it made before.

 

To reach its initial margins of 8%, the agency must either:

 

Reduce workload = streamlining the ideation process
 

Well, that has already happened, given the time limitations on pitches and live briefs today. Trying to squeeze out more efficiency from the creative process is likely to result in more mediocre ideas and more unhappy, ultimately burned-out creative teams.

 

Charge the client more
 

Convincing a client to pay 20% more for big-idea work leads to the problem the industry has been building over the last decades: the value perception of ad agencies.

 

It is unlikely that a client who has just taken half of BAU work in-house to leverage on the cost advantages, is willing to pay the agency 20% more for the remaining big-idea work.

 

After all, if anything, the agency should be happy; now that the boring work is reduced, they can focus on the fun stuff.

 

The key question here is: what is an idea worth? Is it the manhours or the FTEs that have been contractually agreed? Or the expertise? Or the creativity? Or is the value of an idea measured by the results that this idea generates for the business?

 

You know the answer: commercial creativity. It’s just that this industry has not cared much about making the tangible effects of its work to the client’s business visible.

 

To charge more for an idea, an agency will have to determine the value of an idea to the client’s business—so they can ask for a fair share of that value. That requires a deep understanding of the client’s market and business problem, their financials, the competitive environment and so on and agree on metrics that fairly establishes the idea’s contribution to the business.

 

In other words: a new business model.

 

Sooner than you think
 

Given the speed of adoption of gen AI overall (who hasn’t played with ChatGTP yet?), this transformation is likely to happen within the next two or three years. Some industries will move faster (such as retail and FMCG) while others slower (luxury and pharma). We can see the impact already: layoffs in adland have become more common, mainly because of budgets shifts.

 

The biggest driver will be client readiness for technological adoption. The digital transformation in marketing started slow, but accelerated in areas where only budget shifts—without process adjustments—were required (eg SEO, paid digital, social media content). As gen AI is already integrated in some social and ecommerce platforms, this shift is made fairly easy.

 

Future agency model scenarios

To deal with this soon-to-be reality, agencies have two directions to go: go upstream to help solve business problems or downstream and compete on faster/cheaper/easier asset development. This can be delivered through different agency models. Here is a look into the crystal bowl of what might be around the corner:

 

Going upstream

1. The creative marketing consultancy

Concept: solving business problems through strategic and creative marketing thinking.

Service: highly customised to the client’s needs, delivered by a small team of senior, experienced marketeers. Advisory to the CEO and CMO, overseeing and directing marketing initiatives and occasionally executing them.

Remuneration model: value-based; a mix of retainer fee and performance fee, likely a percentage of the results achieved for the business.

Margins: high, depending on the quality of the marketing solutions.

Challenges: progressive clients open to new ways of working and believe in the value of creative problem solving. Access to problem-owners who are aware of the cost of the problem/value of the solution. A measurement framework to validate results achieved.

Variations:

The data-driven agency: a subset of the creative marketing consultancy, utilising data to derive to strategic and creative solutions.

The boutique creative agency: working like in the 90s; the traditional process of advertising development, but on global creative quality level.

 

Going downstream

2. The creative idea supplier

Concept: Cover what gen AI cannot do yet: creative ideas.

Service: A standardized process to develop creative toolkits so clients can develop assets on their own. Toolkit contains idea description, key visuals and key prompts. High integration with gen AI tools.

Remuneration model: package price for toolkit.

Margins: high if standardized processes have been established and are adhered to. Challenges: process discipline. Requires clients who are comfortable to develop and implement assets on their own and are confident in their gen AI skills. 

 

Variations:

The nuclear creative agency – a small team (2-3 people) functioning as a outsourced fractional in-house creative team.

The everything-else agency – develops all assets that the client is not comfortable with generating through gen AI. 

 

Summary:

Whichever way the future will pan out, technological change will dramatically impact the ad industry, as it has in the past. Gen AI is likely to create the most profound impact on the advertising business model yet. So, besides pondering on gen AI’s impact on the creative product, it’s time to think about the business model as well... as time moves faster than we might think.

---

Andreas Moellmann is an independent brand and marketing consultant