The Real Reason D2C Growth Stagnates Even After Bringing Advertising In-House: 'Operational Automation' and 'Structural Design' Are Different Things

The Real Reason D2C Growth Stagnates Even After Bringing Advertising In-House: 'Operational Automation' and 'Structural Design' Are Different Things

Introduction: The Quietly Accumulating Losses in the Shadow of the In-House Boom

Currently, the trend of 'bringing advertising in-house' is accelerating among D2C brand executives and managers.

The series 'The Future of Online Advertising: Transformed by AI In-House Implementation,' reported by Nikkei Cross Trend in May 2026, also details how company after company is breaking away from agency dependence to bring ad operations in-house. Companies that took the plunge into in-house operations following the 'Rakuten Shock,' the case of Orbis reducing CPA by 20%... All of these sound appealing.

However, I want people to pause and think about this trend.

'Bringing ad operations in-house' and 'designing a structure for business growth' are completely different things.

From my perspective as a specialist in LP structural design and CVR improvement, I will dissect this issue today.

Accurately Understanding 'What Was Automated' Through In-House Implementation

First, let's organize the facts.

With media-side AI such as Google's Performance Max campaigns and Meta's automated bidding, 'bid adjustment, budget allocation, and delivery optimization' are almost entirely automated. This is true.

That is precisely why the idea that 'we can cut the management fees paid to agencies' arises.

However, what has been automated is strictly limited to 'optimizing toward set goals.'

AI does not think about 'what KPIs should be set.' It does not judge 'whether that customer has a high LTV.' It does not tell you 'the reason why they do not convert after reaching the LP.'

In the Nikkei Cross Trend article, Mr. Kikuchi of Shirofune also points this out: 'If you leave it entirely to P-MAX, the number of leads increases, but situations where you only gather segments outside the target audience occur frequently.'

If you input 'wrong KPIs' into AI, it will rapidly amplify the wrong results.

This is a phenomenon I have witnessed repeatedly when performing funnel analysis for D2C brands.

The 'Three Structural Problems' Inherent to D2C

D2C brands have structural difficulties that differ from general lead-generation businesses. This is often overlooked in discussions about bringing operations in-house.

1. 'CPA Optimization' and 'LTV Maximization' are in conflict

What is truly important in D2C is not the initial purchase CPA, but the subscription conversion rate, retention rate, and LTV.

However, advertising AI optimizes for 'maximizing initial conversions'.

What happens as a result? You end up attracting a large number of customers who buy only once and then leave, leading to a situation where CPA appears to improve on the surface, but no profit is generated.

This is not an operational issue, but rather a problem of KPI design and profit structure.

Even if you bring operations in-house, if you lack people capable of this design, the AI will faithfully continue to run in the 'wrong direction'.

② 'Mass production' of creatives and 'appeal design' are different things

Generative AI has certainly made the mass production of banners and ad copy easier.

However, what is needed for creatives to produce results in D2C is not 'quantity' but 'precision of appeal'.

• Which customer segment should this product be targeted to,

• what pain point should be used to hook them,

• what emotion should be used to move them,

• and which landing page should they be directed to?

Creatives mass-produced without this design simply consume advertising budgets without leading to results.

'AI can create' and 'can design a selling appeal' are completely different skills.

③ The disconnect between ads and landing pages is the biggest source of loss

This is the point I want to emphasize the most.

What many D2C brands overlook is the problem that 'people gathered through ads are leaving at the landing page'.

Even if you become able to acquire traffic by bringing advertising in-house, if the structure of the landing page does not align with the ad appeal, the conversion rate will not increase.

Looking at data in GA4 or Microsoft Clarity, this pattern can be confirmed in many D2C brands.

• The 'value' appealed in the first view and the 'explanation' in the landing page body are misaligned

• The ad appeals to 'solving a problem,' but the landing page starts with an 'explanation of ingredients'

• The scroll rate is high, but the click-through rate near the conversion button is extremely low.

This is entirely a 'structural design problem' and cannot be solved no matter how much you optimize ad operations.

Advertising is the job of 'gathering.' Landing pages are the job of 'converting.' These two must be designed separately.

Areas that should truly be brought in-house versus areas that should be outsourced

So, how should you think about this?

In my view, there is a clear distinction between the areas that D2C brands should bring in-house and the areas where they should utilize external experts.

Areas to bring in-house

• Daily report checking and numerical tracking

• Basic reading of ad management dashboards

• Collecting customer feedback and accumulating qualitative information

• Managing brand worldview and tone

Areas to partner with external experts

• KPI design and breakdown of profit structures

• LP structural design and construction of CVR improvement hypotheses

• Designing the consistency of messaging between ads and LPs

• Behavioral data analysis using GA4 and Clarity

• Identifying bottlenecks across the entire funnel

In short, 'operations' can be brought in-house. However, 'structural design' requires specialized expertise.

The role of a strategist: Filling the 'void that AI cannot fill'

The most important point made in the Nikkei Cross Trend article is that 'the analytical domain has not yet been fully automated by AI.'

Breaking down your business model, defining long-term metrics to increase LTV, and deciding which metrics to use for optimization—these remain tasks for humans.

As AI accelerates and automates processes, the value of humans who can design 'what to optimize' increases.

This is precisely the area I provide to D2C brands.

• LP conversion design calculated backward from the profit structure

• Consistency checks between ad messaging and LP structure

• Identifying bottlenecks using behavioral data from GA4 and Clarity

• Formulating hypotheses for CVR improvement across the entire funnel

I am involved not as an ad operator, but as a 'strategist who designs structures for business growth.'

Summary: In-house operation is a 'means,' not an 'end'

Bringing advertising in-house is the right direction in terms of cost reduction and knowledge accumulation.

However, bringing it in-house does not automatically mean the business will grow.

Without structural design—such as 'what to acquire,' 'who to convert,' and 'how to maximize LTV'—optimizing operations alone will not accumulate profits.

The more you advance in-house operations, the more you need 'structural designers' rather than 'operators' from the outside.

I believe that whether or not you have this perspective is what separates growing D2C brands from those that do not.

Conclusion

AI moves broadly and quickly. But it cannot decipher the unique structure of your brand.

At NextSTUDIO, we specialize in creating blueprints for improvement by breaking down 'why it isn't selling' based on the brand's profit structure, specifically in LP structural design, CVR improvement, and funnel analysis.

Instead of the broad and fast analysis AI can provide, we spend time deciphering the unique structure of your brand.

I am the only one in charge—that is why I limit the number of projects and dive deep into each company. While large support firms handle multiple companies in parallel, I promise to deeply decipher the unique structure of your brand.

If you are interested, please feel free to consult with me.

👉 NextSTUDIO | LP Structural Design & CVR Improvement

*References to Nikkei Cross Trend in this article are based on publicly available information, combined with the author's own perspectives.