An Evidence Scorecard for Comparing Indiana Agency Proposals

We'd write a short decision brief before scoring anything. Describe the customer you want to serve, the problem you can already observe, and the business outcome that would make improvement worthwhile. Include the current website, your real service area, available staff capacity, and the records you can share. Separate facts from beliefs. “Our office receives inquiries outside our coverage area” is a fact if your inquiry log supports it. “Our competitors have better SEO” remains a hypothesis until someone checks.

Give each candidate the same brief and the same opportunity to ask questions. Otherwise, one proposal may address a lead routing problem while another proposes a broad brand campaign. Those might both be reasonable ideas, but they are not yet comparable. Ask each agency to explain what it would validate first, what it would postpone, and what finding would cause it to change its recommendation.

We score the support behind a claim and ignore how confidently it's presented. A four level scale is enough. Zero means the proposal contains no evidence. One means it gives a general explanation. Two means it provides a relevant example with limitations. Three means it connects relevant evidence to your situation and explains how the work would be checked. Keep a separate notes column so a number never replaces the reasoning.

You can give more attention to the areas that matter most to your project. A website rebuild needs strong migration and handoff evidence. A local campaign needs credible service area targeting and lead handling. Record those priorities before reading the proposals, so the most polished presentation does not quietly change the rules.

Select one important claim from each proposal and ask the agency to walk through it. For a search claim, request the query, location context, date, and observed result. For a speed claim, ask which page was tested and under what conditions. For a lead claim, ask how a lead was defined and which period was compared. Here's how we state one of ours: our Indiana water restoration microsites produced 17 calls and form inquiries from 204 sites during July and August 2026. That count excludes missed calls and doesn't measure jobs or revenue. You are checking whether the team can explain its own work clearly.

A screenshot can be useful, but it rarely tells the whole story. It may show a temporary result, an unusually strong page, or a metric whose definition changed. Ask what the screenshot does not establish. A thoughtful answer is useful evidence of judgment. An agency that distinguishes observation from interpretation makes it easier for your team to evaluate progress later.

Google's guidance on hiring an SEO recommends asking about experience, measurement, communication, and realistic improvements. It also cautions against guaranteed top rankings. Use that guidance as a baseline for the search portion of a proposal, while keeping your own business problem at the center of the review.

A relevant case study should help you understand what changed and why. Look for a starting problem, a defined intervention, a measurement window, and the limitations of the outcome. A business with one location and a small office team may learn more from a modest, clearly documented improvement than from a large revenue figure attached to a very different company.

When reviewing project case studies, ask which parts of the result depended on the client's team. Did the client add staff, change prices, expand availability, or launch a new service at the same time? Those details do not make the marketing work less valuable. They make the lesson more accurate. Your score should reward a clear explanation of contribution, including factors outside the agency's control.

Ask for a specimen report using anonymized or demonstration data. Read it as if you had fifteen minutes before an operating meeting. Can you identify the most important change, the likely explanation, the next action, and the information still missing? If you cannot, request a simpler example. Reporting should make the work easier to manage.

Keep website activity separate from business outcomes. Google's explanation of GA4 engagement rate describes a session based measure. That is different from a qualified inquiry or completed job. An agency can report both, but the proposal should explain the connection it intends to investigate instead of treating one as proof of the other.

We'd suggest a paid discovery exercise only when the project is complex enough to justify one. Agree on a small question, a defined input set, and a concrete output. For example, ask the team to review one service page and its inquiry path, then propose a prioritized improvement plan. The exercise should be useful on its own and should not depend on approving a larger engagement afterward.

A hypothetical business might ask two finalists to diagnose why suitable prospects abandon its quote request. The stronger output could identify a confusing service area statement, a form error, and an unclear response promise. A long keyword list might be less useful for that specific question, even if it required substantial effort. Our contact friction audit workflow provides a practical way to define such an exercise.

Have two people score independently if possible: one who understands the customer and one who will manage the work. Compare their notes before averaging any scores. A disagreement often reveals an unstated requirement. The owner may value speed of delivery while the office manager needs a system that can be maintained without technical help. Both concerns belong in the final decision.

Do not treat a small difference in totals as scientific precision. A scorecard organizes judgment; it does not calculate the correct agency. Record any critical unresolved issue separately, such as an unclear account handoff or a recommendation built on an unverified market estimate. For claims about local demand, use the questions in our market study validation guide before giving the proposal credit.

Write one page explaining the selected next step. Name the business question, the strongest supporting evidence, the expected deliverables, and the assumptions that still need testing. State who will review the work and what information that person needs. This note becomes the starting point for the engagement rather than a record that disappears after the sales conversation.

You should be able to explain the choice without repeating the agency's slogans. The useful outcome is a team you can work with, a scope you can inspect, and a measurement plan that helps you learn. Keep the scorecard and revisit it at the first review meeting. The comparison becomes more valuable when the promises made during selection remain visible during delivery.