Inhousing & One Brain Per Function

The swing that never stops swinging

In-housing is frequently described as a trend. But when you’ve been around the industry as much as I have, you learn pretty quickly that it’s more of a cycle, and one with a surprisingly long history.

The ANA (Association of National Advertisers) has been tracking the phenomenon since 2008. That year, 42 percent of their members had some form of in-house agency.

By 2013 it was 58 per cent. By 2018, 78 per cent. By 2023, 82 per cent. The prediction is that penetration will eventually peak somewhere around 85 to 90 per cent.

But those numbers tell one story.

Here’s another:

92 percent of those same respondents reported still working with an external agency. Which means that the vast majority of companies who have built in-house capabilities have not eliminated the need for agencies. They’ve simply moved certain functions inside while retaining agencies for others.

This cycle (in-house, then back to agency, then in-house again) has been repeating since at least the 1930s. And every three to five years, the pendulum swings. And each time it swings, the arguments on both sides are presented as though they’ve never been made before.

For those of you who are new to this, the arguments for in-housing are about speed, cost control, access to first-party data, brand knowledge, and always-on availability. And those are real advantages. I don’t dispute any of them.

But there’s something that rarely makes it into the in-housing business case. And it has to do with what happens inside a brain that operates on its own. But before we get to that, let’s cover the basic issues.

The homework problem

When you hire one person to run a function, you get one set of experiences, one set of biases, one set of habits, and one ceiling of competence. You also get something that I think is considerably more dangerous: you get a person whose work is never challenged by someone who knows enough to challenge it.

In a company that has spent considerable effort building an in-house marketing team to replace the expense and inconvenience of an agency, nobody has stopped to ask what was lost in the transaction.

Imagine a brand with an in-house team of five. An SEO person, paid media person, content person, creative person, and a marketing coordinator who does a bit of everything. On paper, it’s a well-structured department.

The problem is that each function is covered by exactly one brain.

The SEO person has one way of thinking about SEO. It may be good or mediocre, and nobody in the building can tell the difference because nobody in the building knows enough about SEO to tell the difference.

The same thing with the paid media person who runs the campaigns the way they’ve always run campaigns. There’s nobody sitting beside them who has just come from a different client in a different category with a different set of problems, saying: have you considered this?

Likewise, the creative or the content person writes in one style, because they have one style, and there’s nobody in the next room who writes differently and whose different approach might provoke a useful argument about which direction would actually work.

In an agency, on the other hand, an SEO strategist sits beside four or five other SEO strategists who have all come from different clients and made different mistakes. They’ve all seen different things work and different things fail. So when one of them proposes an approach, the others can say: we tried that on a retail client last year, and it didn’t work because of X. Or: here’s something we did for a financial services brand that might be relevant.

It’s the reason that five radiologists reading the same scan will catch things that one radiologist misses. It’s also the reason that engineering teams do code reviews and why any serious creative process involves showing work to people who weren’t involved in making it.

And an in-house marketer, no matter how talented, is a single radiologist reading every scan alone. They may be excellent. They may catch 90 percent of what there is to catch. But the 10 percent they miss has no mechanism for correction, because nobody in the building knows enough to see what they’ve missed.

I think there’s an element of the Dunning-Kruger effect at work here: the cognitive bias where people with limited competence in a given area tend to overestimate their own ability, precisely because they lack the knowledge to recognise what they don’t know. An in-house team that operates in isolation has no external reference point against which to calibrate its own performance. They may be doing a good job. But they may also be doing the wrong job or the right job the wrong way.

With that understanding in place, let’s now take a look at the arguments for In-house marketing, beginning with the argument about speed.

The speed problem (which is really a pressure problem)

Yes, in-house teams can move faster than agencies, which are encumbered by briefing processes, approval chains, and the general overhead of client service.

There’s truth in that. But in my experience, it’s incomplete truth.

In-house teams can move fast when they choose to. But they are also, by their nature, operating without the external pressure that forces agencies to deliver.

An in-house team answers to one boss in one company. If the content is late, it’s late. If the campaign slips a week, it slips a week. There’s nobody banging on the door saying: we promised the client this by Thursday. The urgency is self-generated, and the problem with self-generated urgency is that it tends to dissipate over time.

An agency has deadlines imposed by clients. It has other clients competing for the same team’s attention. And it has the ever-present awareness that if it doesn’t deliver, it can be replaced. These are not just part of the job. These are not comfortable conditions, but they are conditions that produce output.

Now let’s look at the argument for cost-control.

What agencies actually sell (and why it’s hard to put on a spreadsheet)

Yes, going in-house could, conceivably, save you money. But at what cost to other things?

There was once a time, so the old hands tell me, when client companies would welcome an agency’s thoughts on just about all aspects of their business: diversification, brand strategy, investment, internal training, presentation, as well as advertising and promotion. For a wide variety of reasons, all that has changed. Increasingly, clients expect only creativity and strategy from their agencies, and increasingly, that’s all that agencies are organised to provide.

I think this narrowing is part of what’s driving the in-housing cycle. Companies look at what their agency does (makes ads, buys media, writes copy) and quite reasonably conclude that they could hire people to do those things internally. And they can. The executional work can often be done in-house for less money and more quickly.

But what gets lost is the thing that’s hardest to see on an invoice: things like an outside perspective or the accumulated cross-client intelligence that an agency acquires by working across multiple brands in multiple categories or the insights that come from having seen your problem before or the challenges and pushback that come from people who don’t work for you and who therefore you unconsciously give more credibility to as experts.

All of these are real benefits of hiring outside agencies.

Unfortunately they are also psychological truths that are hard to admit (that we tend to trust people outside of close relationships more than people in them) and not things that are easy to put in line items.

Which brings me back to the McNamara fallacy that I discussed a few issues ago:

The fallacy that causes you to only measure what you can measure, dismiss what you can’t, presume that what you can’t measure isn’t important, and conclude that what you can’t measure doesn’t exist.

The cost of an in-house team is easy to measure.

The cost of what you’ve lost by eliminating the agency is not.

And so, by the logic of every spreadsheet ever created, the in-house team is going to win because it’s cheaper.

The thing you need to ask yourself is, cheaper at what?

Which brings us to the argument about access to first-party data and brand knowledge.

The IKEA effect in reverse

There is no question that the marketing team that works in-house is closer to the data and the brand than an outside agency.

Not only that, but when an in-house team builds something (a campaign, a website, a brand strategy, a media plan), the people who built it are personally invested in its success. The problem with this, of course, is the age-old argument that the worst person to create an advertisement is the person closest to the brand. Even Steve Jobs had the wisdom to hire an ad agency, Chiat Day, to create the ad campaigns that put Apple on the map.

But there is another problem when the people closest to the brand are also creating the brand, and it has to do with a cognitive bias called the IKEA effect. For those of you who are not familiar with it or have never assembled IKEA furniture, the IKEA effect states that you value things more when you’ve helped create them. In a healthy creative process, this investment is a good thing. It produces ownership, care, and attention to detail.

But in an in-house context, where there’s no external perspective to challenge the work, the IKEA effect can become a trap. The team has built something, and they believe in it. Their boss (who hired them and needs their hire to be a good decision) also believes in it. And there’s nobody in the room with the distance, the independence, and the cross-category experience to say: I think this might be wrong.

I’ve seen this pattern repeatedly. An in-house team builds a campaign. Results are disappointing. The team attributes the disappointing results to external factors (the market, the competition, the economy). And they can get away with it because no one within the organisation has the expertise to diagnose the real problem. And so the campaign continues, or a new idea is hatched and results continue to disappoint. Eventually, the call and some agency is brought back in. The cycle begins again.

The hybrid truth

I suspect the answer (as with most things in this industry) is not either/or. The ANA data confirms this: 82 per cent of companies have in-house capabilities, and 92 per cent still use external agencies. The question is how you structure the hybrid.

The companies I’ve seen do this well tend to follow a pattern. They keep in-house the work that benefits from proximity to the business: content production, day-to-day digital management, social media monitoring, the fast-turnaround executional work that agencies are too slow and too expensive to do efficiently. This is the commodity layer. And AI is increasingly doing much of the heavy lifting here.

They keep with agencies the work that benefits from multiple perspectives: strategic thinking, creative development, media strategy, and (perhaps most importantly) the diagnostic function. The ability to look at a declining business and say: the problem isn’t the campaign. The problem is the product. The positioning. And something you can’t see because you’re inside it.

And that diagnostic function is the most valuable thing an agency provides. It’s the first thing that disappears when a company fully in-houses its marketing. Because diagnosing problems requires distance. It requires the ability to compare what you’re seeing in one business with what you’ve seen in twenty others. And it requires the willingness to say uncomfortable things to people who sign your invoices but don’t control your employment.

An in-house marketer who tells the CEO that the product is the problem is risking their job. An agency that tells the CEO the product is the problem is risking a client. One of those is a career-ending conversation. The other is a Tuesday.

Three questions for the company that’s considering in-housing

First: for each function you’re planning to bring in-house, ask yourself who will mark the homework. If the answer is nobody (because nobody else in the building has the expertise to evaluate the work), you have a structural blind spot that no amount of talent can fix.

Second: look at the last three major marketing decisions your company made. How many of them were challenged by someone who had no personal stake in the outcome? If the answer is zero, consider the possibility that your marketing is operating in an echo chamber, and that the absence of external friction is costing you more than the agency retainer ever did.

Third: ask yourself honestly whether you’re in-housing because the work will be better, or because it will be cheaper and more controllable. Both are legitimate motivations. But only one of them will appear on the spreadsheet, and the spreadsheet, as we have discussed at some length in this series, is a notoriously unreliable narrator.

Remember: The cycle will keep swinging.

It always has.

Companies will bring marketing in-house, discover what they’ve lost, bring agencies back, forget what they gained, and bring marketing in-house again. The question is whether you want to learn this lesson from your own experience (which takes about three years and costs roughly whatever your competitors gained while you were figuring it out) or from someone else’s.

I know which one I’d recommend.

But then, I would say that, wouldn’t I?

Hit reply if you have thoughts. I’m listening.

Jonathan