Every agency switch starts the same way. You're frustrated. Performance is flat or declining. Your account manager can't explain why last month was bad. Maybe they keep sending you the same cookie-cutter report with a paragraph that basically says "we're monitoring closely" — which, in my experience, means nobody's actually looked at the account in two weeks.
So you start shopping around. You take a few calls. The new agency shows you a shiny audit with all the things your current team is doing wrong. And you think: finally, someone who gets it.
I've been on both sides of this. I've been the new agency doing the audit, and I've been the one watching a client leave. And I need to tell you something that most agencies won't: switching costs way more than you think it does. Not in fees. In performance, time, and knowledge you didn't even realize you had.
The 60-90 Day Performance Dip Is Real
Nobody talks about this in the sales process. But here's what happens the moment you hand your Google Ads or Meta account to a new team.
They don't know your account yet. They don't know which campaigns were intentionally structured that way because of a test you ran eight months ago. They don't know that you paused that keyword group in March because it drove a bunch of leads that your sales team said were garbage. They don't know that your best-performing audience on Meta was one your old team built from a customer list you uploaded during a promotion last Black Friday.
So they start making changes based on what they see right now, without the context of what came before. And some of those changes are improvements — sure. But some of them undo months of learning.
I onboarded a client last year who'd been with their previous agency for three years. Good agency, actually. The client just wanted a change of pace, fresh eyes. Fair enough. But in the first 60 days, their cost per lead went up 35%. Not because we were doing anything wrong. Because their old campaigns had been optimized over three years of accumulated data and we essentially had to rebuild that intelligence from scratch.
We recovered by month three and beat their old benchmarks by month four. But that 60-day dip? That was real money. About $18,000 in additional cost-per-lead overhead that nobody budgeted for.
Conversion Tracking Gaps Will Bite You
Here's the thing that keeps me up at night when we take over a new account. Conversion tracking.
Every agency sets up tracking slightly differently. Maybe your old team was using Google Tag Manager with a specific container structure. Maybe they had custom events firing on certain pages. Maybe — and this happens more than you'd think — their tracking was broken in some subtle way that inflated the numbers they were reporting to you, and when the new agency sets it up correctly, the "performance drop" is actually just accurate data for the first time.
That's a fun conversation. "Hey, so your leads didn't actually decrease. Your old agency was double-counting conversions from their thank-you page redirect. You were never getting 200 leads a month. It was closer to 140."
I've had that exact conversation three times this year alone. Three times.
During the transition, there's almost always a gap where tracking isn't fully set up or validated. Could be a few days, could be a couple weeks if the implementation is complex. During that window, you're flying blind. Spending money on ads with no clear picture of what's working. It's like driving at night with your headlights off. You'll probably be fine. Probably.
The Institutional Knowledge Problem
This is the big one. The one that nobody puts a dollar figure on because it's hard to quantify. But it might be the most expensive cost of all.
Your current agency — assuming they've been around for a while — knows things about your business that aren't written down anywhere. They know that your CEO hates seeing the brand name in ad copy (learned that the hard way in Q2 2023). They know that your best customers come from searches related to a problem your product solves, not the product name itself. They know that you tried YouTube ads in 2022 and it was a disaster, but the real reason it failed was because your landing page wasn't ready, not because the channel doesn't work.
All of that knowledge? It lives in someone's head. Maybe in a Slack thread from 14 months ago. Maybe in a meeting note that nobody will ever look at again. When that agency-client relationship ends, that knowledge walks out the door.
The new agency starts fresh. And they're going to spend the first three months re-learning things your old team already knew. They'll probably suggest YouTube ads. They'll probably put the brand name in the copy. They'll probably test some things that were already tested. Not because they're bad at their jobs — because they literally don't have access to the history.
I've seen clients who've switched agencies three times in five years. Each time, the new agency runs the same tests, makes the same discoveries, and arrives at roughly the same strategy as the previous one. Eighteen months of "learning" that could have been three months if the knowledge had transferred properly.
What It Actually Adds Up To
Let me put some rough numbers on this for a mid-market B2B company spending $30K/month on Google Ads.
- Performance dip (60-90 days): 20-40% increase in cost-per-lead during the transition period. If your normal CPL is $150, you're looking at $180-$210 for two to three months. That's $9,000-$18,000 in extra acquisition cost.
- Tracking gap (1-3 weeks): Potentially flying blind on $7,500-$22,500 in ad spend during setup and validation.
- Onboarding time: Your marketing team is going to spend 10-20 hours in calls, emails, and Slack messages getting the new agency up to speed. That's real internal cost.
- Knowledge rebuilding: The hardest to quantify but arguably the largest. Every insight your old agency accumulated — which audiences respond, which messaging resonates, which days of week perform best for your business — has to be re-learned through new spend and new tests.
I'd estimate the total hidden cost of switching, for a mid-market account, is somewhere between $30,000 and $60,000. That's on top of whatever you're paying either agency in fees.
OK So When Is It Actually Worth It?
I just spent 800 words telling you how expensive switching is. Now let me tell you when you should do it anyway.
When performance has been declining for 6+ months and your agency can't explain why. A bad quarter happens. Two bad quarters is a trend. If you've raised the issue and all you get is "the market is competitive" without a concrete plan to fix it, leave.
When your strategist keeps changing. If you're on your third account manager in two years, the knowledge problem I described above is already happening inside your current agency. You're paying the switching cost without even switching.
When you've outgrown the team. Some agencies are great for companies spending $10K/month but don't have the expertise for $100K/month. That's not a criticism — it's a specialization issue. If your business has grown and your agency hasn't grown with you, it's time.
When you catch them being dishonest. Inflated numbers, hidden fees, taking credit for organic conversions in their paid reports. I've seen all of it. If trust is broken, no amount of institutional knowledge is worth staying.
When the agency is clearly just going through the motions. You know the feeling. Same report, same recommendations, no new ideas. They've stopped being curious about your business. At that point, the switching cost is less than the cost of stagnation.
What If You Didn't Have to Start From Scratch?
Here's what I keep coming back to after ten years of doing this work. The reason switching is so painful isn't really about the technical stuff — the tracking setup, the campaign restructuring, the bidding strategies. A good team can figure all of that out in a few weeks.
The real pain is the knowledge loss. It's the fact that everything your previous agency learned about your business, your customers, your competitive dynamics, your past tests — all of it disappears when the relationship ends. And the next agency starts from zero.
That's a structural problem with how most agencies operate. Knowledge lives in people's heads, in Slack threads that get archived, in meeting notes that nobody indexes. There's no system for making sure that intelligence compounds over time instead of resetting every 18 months.
I think about this a lot. What would it look like if every insight, every test result, every piece of competitive intelligence, every client preference was captured in a way that survived team changes and agency transitions? What if the knowledge about your business only ever grew, and never got lost?
That's a solvable problem. And I'd argue it's the most important problem in the agency-client relationship that nobody is seriously trying to solve.
But that's a topic for another post.