THINXSTER
Blog/AI Agency
AI Agency8 min readJune 11, 2026

How to Compare Marketing Agencies: A Scoring Framework That Cuts Through Pitches

Stop comparing agencies on portfolios and vibes. Use this 5-factor scorecard — outcomes, transparency, systems, speed, incentives — to pick the right partner.

RK
Ryan Korsz
Founder & CEO, Thinxster

TL;DR

Stop comparing agencies on portfolios and vibes. Use this 5-factor scorecard — outcomes, transparency, systems, speed, incentives — to pick the right partner.

→ See how this applies to your business (free 30-min call)

Every agency pitch deck looks the same: logos of past clients, a couple of cherry-picked case studies, a services grid, and a team photo. Which is exactly why comparing agencies by their pitches doesn't work — you're comparing their marketing of themselves, and the agency best at marketing itself is not necessarily best at marketing you.

After watching hundreds of businesses go through this decision (and being one of the options on the shortlist), here's the comparison framework I'd use if I were on your side of the table. Five factors, scored 1–5 each, with the exact questions that expose the score.

First, Know What You're Actually Comparing

Agencies cluster into four models, and comparing across models without realizing it is how businesses end up disappointed:

  • Traditional full-service — retainer-based, account managers, monthly reporting. Buys you breadth and hand-holding; usually slow and activity-priced.
  • Niche specialists — one channel (just Google Ads, just SEO) or one vertical (just dentists). Deep expertise, but you'll own the gaps between specialists.
  • Performance/pay-per-lead shops — you pay for leads delivered. Aligned-sounding, but quality control becomes the battleground, and they often sell the same leads to your competitors.
  • AI-native agencies — built around automated systems (AI lead response, automated nurture, closed-loop attribution) with humans on strategy and creative. Faster and more measurable; the catch is that plenty of traditional shops have stapled "AI" onto their homepage without changing how they operate, so this category needs the hardest verification.
  • Decide which model fits your stage first. Then compare within and across using the scorecard.

    Factor 1: Outcomes (Not Case Studies)

    A case study is a story. An outcome is a number with a methodology. The difference shows up under questioning.

    Ask every agency: "For a client with roughly my ticket size and budget, what was the cost per acquired customer, and how did you measure it?" Then follow with: "What's a client engagement that didn't work, and why?"

    The first question tests whether they track to revenue at all. The second tests honesty — every real agency has failures, and the ones who claim otherwise are telling you how they'll handle your bad months: with spin.

    Score 5 if they volunteer revenue-level numbers with the measurement method attached. Score 1 if everything is reach, impressions, and adjectives.

    Factor 2: Transparency

    You're testing one thing: will you be able to see what's happening without asking permission?

  • Do you keep ownership of your ad accounts, CRM, and data? (Non-negotiable. If the answer is no, stop scoring and remove them from the list.)
  • Is there a live dashboard, or do numbers arrive monthly as a PDF?
  • Can you see spend, not just results? Some agencies bundle ad spend into their fee precisely so you can't compute the markup.
  • The quality of an agency is inversely proportional to how hard it is to see your own numbers.

    Score 5 for live dashboards on accounts you own. Score 1 for monthly PDFs from accounts they control.

    Factor 3: Systems

    This factor separates agencies that produce activity from agencies that produce infrastructure. Activity stops when the retainer stops. Infrastructure keeps working.

    Ask: "Walk me through what happens in the first ten minutes after a lead comes in from one of your campaigns." The answer tells you everything. Weak agencies describe ads and say lead handling is your job. Strong ones describe a machine: instant first contact, qualification, booking, CRM write-back, nurture for the not-yet-ready.

    90s
    the lead response time a modern agency should be building toward — Thinxster's AI callers hit it around the clock

    Also ask what happens to leads that don't answer the first call. The honest industry answer is "usually nothing" — most businesses follow up once or twice and quit. An agency with real systems has an automated multi-touch sequence and can show you its conversion rate.

    Score 5 for a concrete, automated answer with numbers. Score 1 for "we hand the leads to your team."

    Factor 4: Speed

    Speed shows up in two places: how fast their systems respond to your leads (covered above) and how fast they ship work. Ask for their testing cadence — how often do new creatives, audiences, or landing pages go live? Ask how long onboarding takes before the first campaign is live, and what the first two weeks look like.

    Good answers are specific and short: tracking wired in week one, first campaigns live in week two, new tests shipping weekly. Vague answers ("we move as fast as the strategy allows") usually mean a queue behind other clients.

    Score 5 for documented weekly iteration. Score 1 for quarterly "strategy refreshes."

    Factor 5: Incentives

    Finally, look at how the agency makes money, because their pricing model is a prediction of their behavior.

  • Flat retainer: stable, but the incentive is client retention, not growth — comfortable mediocrity is profitable.
  • Percentage of ad spend: the incentive is for you to spend more, whether or not it converts better.
  • Performance components: fees tied partly to outcomes align incentives, but verify how the outcome is counted (a "lead" definition can be gamed; booked revenue can't).
  • There's no perfect model, but there's a perfect test: propose adding an outcome-linked component and watch the reaction. Confidence welcomes accountability; theater resists it.

    The Reference Call That Actually Tells You Something

    Every agency will hand you references, and every reference will be friendly — that's why they were chosen. The trick is asking questions the agency didn't prep them for:

  • "What broke in the first 90 days, and how did they handle it?" Every engagement has early friction. You're listening for whether the agency surfaced problems proactively or got caught.
  • "How long does it take to get an answer when you ask a numbers question?" This verifies the transparency score from the source.
  • "If you could change one thing about working with them, what would it be?" The answer is almost always honest, because it's framed as advice rather than criticism.
  • "Did your cost per customer actually improve, and do you know that from your own books or from their reports?" The most useful question on the list. A reference who knows their CAC from their own accounting is proof the agency builds measurable systems. A reference who only knows what the reports said is proof of good reporting, which is not the same thing.
  • Twenty minutes of this is worth more than every slide in the pitch deck.

    Common Mistakes That Sink the Comparison

    Watch for these in your own process, because they're how good frameworks produce bad picks:

  • Weighting price too early. A $2,500/month agency that produces customers at $300 each is dramatically cheaper than a $1,500/month agency producing them at $700. Score the five factors first; bring price in last, as cost per outcome rather than cost per month.
  • Being charmed by the strategist who won't run your account. Agencies send their best people to pitches. Ask directly: "Who works on my account day to day, and can I meet them before signing?" If the answer is evasive, the A-team is for closing, not delivering.
  • Comparing proposals instead of answers. Proposals are written to be compared favorably. The scorecard works because it's built on live answers to questions they didn't fully script — keep the weight there.
  • Skipping the small test. If you genuinely can't separate two finalists, ask each for a scoped 60–90 day starting engagement with defined success metrics instead of an annual contract. The one who agrees readily has told you something; the one who insists on twelve months has too.
  • Running the Comparison

    Put your shortlist in rows, the five factors in columns, and score every agency from the same conversation script. Two patterns to watch for:

    1.

    A high outcomes score with a low transparency score is a contradiction. If the results are real, why is the data hard to see? Trust the transparency score.

    2.

    Systems beats everything else for service businesses. A mediocre ad manager feeding a great response-and-nurture machine outperforms a brilliant ad manager feeding a void. Weight Factor 3 double if your leads are calls and form fills.

    $102M+
    revenue generated by Thinxster client systems — built on exactly this systems-first weighting

    We built Thinxster to score well on this exact rubric — revenue-level reporting, client-owned accounts, AI response systems, weekly iteration — not because it's fashionable but because it's what we'd demand if we were buying. Hold us to it. Hold everyone to it.

    If you're comparing agencies right now, [book a free strategy call](/book) and bring your shortlist. We'll walk through the scorecard on your actual situation — and if another shop on your list is genuinely the better fit, we'll tell you that too.

    Free Weekly Briefing

    One AI Marketing Tactic.
    Every Tuesday. Free.

    What's actually working across our client accounts right now — ROAS moves, follow-up sequences, creative angles. The stuff that isn't in any blog post yet.

    No spam. Unsubscribe anytime. 1,200+ business owners already in.

    Ready to Deploy

    SEE THIS IN
    YOUR BUSINESS.

    30 minutes. We scope the exact systems that apply to your situation and give you a plan.

    ★★★★★ Trusted by 47+ local service businesses

    BOOK A STRATEGY CALL →