Why I brought paid search in-house.
Paid search had become too important to manage at arm’s length. I wanted the speed, data, economics and learning loop closer to the people responsible for the business outcome. That did not mean every outside partner was wrong. It meant this capability had crossed a line.
Paid search had become a business capability, not just a media channel.
The issue was not that an agency could not run campaigns. The issue was that too much of the learning loop sat outside the business.
As the channel became more important, I cared less about whether the weekly campaign report looked good and more about a different set of questions: How quickly could we act on new information? Could we connect media decisions to downstream conversion? Did we understand the economics ourselves? Where was the knowledge accumulating?
At that point, keeping the capability at arm’s length was creating more friction than leverage.
It came down to six practical questions.
The more important a capability is to growth, the more valuable direct operating knowledge becomes.
Shorter decision loops matter when spend, conversion behavior and market conditions change quickly.
A campaign can look efficient in-platform while producing weak downstream economics.
Bringing work inside adds management load and talent cost. The economics still have to work.
Budget shifts, attribution questions and optimization choices become easier to inspect when ownership is direct.
If the capability will matter for years, the location of the learning matters too.
Moving the work inside only mattered if the way we ran it changed.
I did not want to reproduce the same operating model with a different org chart. The feedback loop had to get tighter, and the definition of performance had to move closer to the business outcome.
Managed at a distance
- Execution and optimization primarily outside the business
- Reporting mediated through an external operating layer
- Platform performance easier to see than downstream economics
- Channel knowledge accumulated with the partner
- Changes required more coordination across organizational boundaries
Capability owned inside
- Direct ownership of campaign execution and optimization
- Direct access to the data and assumptions behind decisions
- Performance connected more closely to conversion and business outcomes
- Budget decisions made closer to the operating context
- Knowledge and learning compounded inside the organization
The dashboard had to answer a business question, not just a media question.
We still needed platform metrics. I just stopped treating them as the final definition of success.
Improvement in trailing-12-month paid-search acquisition efficiency after the capability moved in-house. Just as important, the business could now see and challenge the system directly.
What changed in the weekly view.
The old question was “how is paid search doing?” The better question was “what is this channel producing for the business, and what should we change next?”
I would make the same decision again—but not because in-house is always better.
If the work is central to growth and the organization needs to get smarter at it over time, knowledge location matters.
The data, measurement, workflows and decision rights have to change too.
The useful question is not “agency or in-house?” It is which parts need internal ownership and which benefit from outside specialization.
The closer optimization gets to downstream economics, the less likely the organization is to scale activity that looks good in-platform but produces a weak business result.