How to Standardize Agency Delivery Without Losing the Innovation

Innovation without guardrails kills your margin, and optimization without innovation kills your agency.

I’m a marketer and strategist at heart, and innovation is my favorite part of agency life. I love trying the new tool before anyone else, cracking a client problem in a way nobody’s tried, and watching a team get fired up about the work again. That’s the stuff that wins pitches and keeps your best people around.

It took me a long time to learn what it costs.

I was the strategist who blew every estimate

Early in my career I was the innovative strategist on the delivery team, and I blew past every time estimate I was given. I inflated the cost of everything I touched, because I didn’t know what the work cost and I didn’t care.

The agencies I worked at were lifestyle businesses. For better or for worse, they let me do my thing. They let me cook, and I never gave one thought to margin or profit. It was all about doing cool work for cool clients and getting results. The work was great. The margins were not.

That changed when I started working with agencies that have ownership groups and outside investment. Once somebody is reading the numbers every month, the balance between great work and margin gets crystal clear. Are we getting results for clients? Are we staying within margin? And can we see both while the work is happening, instead of at the end-of-month close?

Knowing the problem never gave me time to fix it

Last year I was running an agency, and all the stuff I’d ignored for years was sitting on our EOS rocks list. Every month after the books closed, our fractional CFO would come in and give us the hard truth. Rightsize payroll. Optimize billable utilization. Put guardrails around overservicing.

I knew he was right, and I also knew I wouldn’t do anything about it. I couldn’t physically do it. I had 6 client calls the next day. Every one of those meetings made it worse, because now I knew exactly what was wrong and I still couldn’t get my head above water long enough to change anything.

That’s the part of the margin conversation that gets skipped. The books tell you something is off a month or more after it started, and by then the person who could fix it is buried in client work.

The pendulum has swung too far the other way

Now everyone’s telling agencies to standardize, productize, and automate everything with AI. And I get it. You NEED standardization to build a scalable delivery model.

But lock it down too tight and you turn into the commodity agency. The work gets predictable, the pitch sounds like everyone else’s, and you end up competing on price. Your best people feel it first. The strategists and creatives who joined to do interesting work get bored, and the good ones leave. The clients who came for that thinking follow them out.

There’s such a thing as too efficient. That’s where you AI and automate the heart and soul right out of the work.

Standardize the core, then budget the experiments

Here’s how I help agencies have it both ways.

Standardize the core. The repeatable stuff should run the same way every time, because that’s what lets you scale. Most of a service is steps the client never sees, and those steps don’t need anyone’s creativity.
Be intentional about where and how you innovate. Use AI and automation on the repeatable work to free up time for the team to try new tech, techniques, and processes. For most junior and mid-level production roles, that’s 5-10% of their time. Senior people get a lot more, and any billable work they touch should be pushing something new forward.
Track delivery health KPIs in real time. If an experiment is eating the budget, you want to know this month, not at the year-end close. This is the piece I was missing when the CFO calls kept coming.
Find out what people are learning on the fly. A lot of the best ideas in an agency are a shortcut somebody found mid-project and never mentioned. Contextual time tracking surfaces where people are running the process differently. Lunch and learns and share sessions give them a place to show off what they figured out.
Vet it, highlight the wins, and fold them back into the system. Check the new way against the numbers to see if it saves time, and against your quality bar with blind deliverable scoring. When it wins, give the person credit in front of the whole team, then update the SOPs, video walkthroughs, and AI skills so everyone works that way next time.

“We can’t afford to give up 10% of billable time”

This is the first pushback I hear, and it’s a fair one. On paper, giving a production team 5-10% of their week to try new things looks like 5-10% of your capacity walking out the door.

The time comes out of the repeatable work you standardized and handed to AI and automation. If the core runs faster, the team has hours they didn’t have before. The only question is whether those hours go to more of the same work or to figuring out the next version of it.

The tracking keeps that honest. If an experiment is eating a client’s budget, you see it that month and you rein it in. If it’s saving time, you see that too, and it goes into the system for everyone. Without the numbers, you find out what all that innovating cost you at the year-end close, which is right where I was sitting with our CFO.

The step agencies drop most often is the last one. Agencies run the experiment, somebody finds a better way, and it stays in that person’s head. The SOP never changes, the next hire learns the old way, and the agency paid for the experiment without keeping the result.

Creativity is the endgame

Creating the space for people to be human, have fun, and put creativity and strategy into the work is the endgame. The guardrails are what let you afford it.