Most Agencies Don't Want you to know this... | Upside #94

Strategy

8 Jul

Why Most Agencies Fail Brands (And How to Actually Vet One)

Most brands who come to us hate agencies, or have been burned by one. This episode breaks down why the standard agency model tends to fail, and gives the actual questions to ask before hiring one, plus the signs it's already going wrong.

In this episode: Ali and Zamir break down why the typical agency model breaks down and what to look for in a genuine growth partner.

Key takeaways

The average brand-agency relationship lasts around 8 months, a sign the model itself is often broken, not just bad luck with one bad agency.
Most agencies need a scalable, formulaic process to make their business work, which means one-size-fits-all delivery rather than the strategic, hands-on approach only a small share of brands actually need.
Payment structures often reward input, not output: percentage-of-ad-spend and flat retainers pay agencies more for spending your money or just keeping the lights on, regardless of your results.
A simple litmus test: describe a product with reviews below 3.8 stars and ask if they can grow it as-is. Any agency with integrity should say no, since no amount of ad spend fixes a fundamental product problem.
Get them to audit your account before you commit, and meet the actual person who'll manage it, not just whoever sold you the contract.
Expect a real learning phase, roughly 3 months, before changes should start showing results. Anything faster is likely knee-jerk, not strategy.
Pick 3 measurable metrics tied to your actual objective, and ask the agency which metrics best judge that objective rather than dictating them yourself.

Timestamps

00:00 — Why brands come in with agency baggage
01:53 — Why the agency model tends to fail (the incentive problem)
07:00 — What you should expect from an agency day-to-day
07:41 — Why input-based pay misaligns incentives
11:22 — How to vet a partner: the poor-review litmus test
14:57 — Why the learning phase takes about 3 months
17:55 — Seven questions to ask your agency right now
20:18 — Red flags and green flags to watch for

The detail

Why the standard agency model tends to fail
Agencies need a scalable, repeatable process to run their business, which means most are built around one-size-fits-all delivery rather than a strategic, hands-on approach. That works fine for brands that just want things kept ticking over, but not for brands chasing significant growth or a specific goal, those need something closer to special forces than general infantry. Layer on how most agencies get paid, a retainer or a percentage of ad spend, both reward input rather than results, and average tenure with an agency sitting around 8 months starts to make sense.

The litmus test for vetting a partner
Describe a real product, or invent one, with a review score below 3.8, and ask whether the agency can grow it as it currently stands. An agency with integrity says no, no amount of budget or optimisation overcomes a genuine product problem, and their willingness to say so tells you whether they'll be straight with you later. Beyond that, ask for an account audit before signing anything, and insist on meeting the actual person who'll manage your account day to day, not just whoever sold you the contract.

Give the relationship a real learning phase
A good partner spends the first few weeks understanding your business, category, and existing account logic before making changes, that learning curve is normal and takes around 3 months to fully show in results. If a partner skips straight to implementation with no ramp-up, that's a sign they're running a cookie-cutter process rather than actually learning your account.

The seven questions to ask your agency right now
What percentage of revenue is organic versus paid, and is that trending the way you want? What's the TACOS trend over the last six months? When was listing copy and images last updated? Is A+ content and brand story set up? What are current stock levels and replenishment timing? What keywords are ranking organically in the top 10? What are the targets for the next six months, and how were they set? Pick at most three of these to track closely, tied to the actual goal, more than that gets unwieldy, fewer risks losing the plot.