I once spent fourteen hours building a monster. It was a vendor-selection spreadsheet with

, each weighted to the second decimal point. I thought I was being rigorous. I thought I was protecting the project from human bias, from the “vibe” of a charismatic salesperson, and from my own fallibility.

I felt like a technician of truth. Looking back at the text messages I sent to my partner during that week-complaining about the “lack of objectivity” in my peers and my obsession with “data-driven selection”-I realize I wasn’t seeking truth at all.

The project did go south. It didn’t just go south; it cratered. The vendor we selected had the highest “Weighted Score” because they ticked every box in the RFP. They had the liability insurance. They had the three case studies in the exact same vertical.

They had a “dedicated QA department” (which turned out to be two interns in a different time zone). But the spreadsheet couldn’t measure the one thing that actually mattered: their inability to communicate a complex technical architectural shift to a non-technical marketing lead.

The spreadsheet was “defensible,” but the outcome was a disaster. I realized then that I had optimized for the one thing that doesn’t help the client: my own safety.

The Humming Room of Procurement

I think about that spreadsheet whenever I see a modern RFP. You know the ones. Forty questions, none of which predict whether the website will actually work on a Tuesday in March.

The room is usually the same. There’s a boardroom table, a projector humming with a slight, dusty wheeze, and four agency columns on a screen. At the bottom, a row shaded in a soft, non-threatening green shows the final scores. Two of the agencies are within 0.3 points of each other.

The procurement officer, sensing the stalemate, adds a new tab labeled “tiebreak.” Having no better instrument to measure the soul of a creative partner or the integrity of a codebase, they sort on price.

Nobody in that room has opened any of the four agencies’ past client sites on a mobile phone to see if the navigation menu breaks. Nobody has checked to see if the “SEO-friendly” sites they built actually have schema markup that an AI assistant can parse. They are too busy verifying the “Sustainability Policy” of a three-person design shop.

This is the central paradox of procurement: it is treated as risk reduction, but it only reduces the risk of blame. If you follow the rubric and the project fails, you aren’t the problem-the rubric was. But if you follow your intuition, hire the unconventional agency that actually understands your business, and it fails, your head is on the block.

Positions vs. Interests

In my work as a mediator, I see this play out in the “Positions vs. Interests” dynamic. In a negotiation, a “Position” is what you say you want: “We want a vendor with fifteen years of experience.” An “Interest” is why you want it: “We are terrified that if the site goes down during our product launch, we won’t know who to call.”

When you write an RFP, you are codifying your Positions. You ask how many years the firm has been operating. You ask if they carry two million in liability. These are easy to score on a five-point scale. They make the procurement officer feel like they have done their due diligence. But these questions are proxies for trust, and they are very poor ones.

A firm can have twenty years of experience and still build a site that requires a developer to change a comma in a headline. A firm can have five million in liability insurance and still deliver a “pixel-perfect” design that degrades into a jumbled mess on a tablet.

The questions that actually predict a successful outcome-questions about CMS architecture, about whether the marketing team will be able to publish a page without filing a ticket, about whether the team that scopes the project is the same team that builds it-are hard to score. They require the evaluator to actually understand the work.

The Invisible Cost of Mediocrity

Selection systems built to protect the selector rather than the project produce mediocre outcomes with excellent documentation. The cost is invisible because the failure doesn’t happen on the day the contract is signed.

It happens eighteen months later, when the marketing team realizes they are locked into a platform they can’t manage, or when the “SEO strategy” turns out to have been a handful of meta tags added as an afterthought.

By then, the procurement team has moved on to the next spreadsheet, and the failure is attributed to “execution” rather than the “tiebreak” tab.

Most of the truly successful digital projects I’ve seen weren’t the result of a “perfect” RFP process. They happened because someone in the organization was brave enough to break the process. They looked at the work. They talked to the actual developers. They insisted on seeing how the CMS worked from the perspective of an editor, not just a visitor. They looked for a partner who was willing to be transparent about their process from day one.

This transparency is rare because it’s dangerous for a mediocre vendor. It’s much easier to hide behind a vague “Scope of Work” than it is to show exactly how a project will be executed.

When you see a line-by-line scope, you aren’t looking at a marketing document. You are looking at a roadmap of accountability. You are seeing exactly which parts of the build are being handled in-house-branding, UX/UI, Webflow development, SEO, and full-stack engineering-and which parts are being ignored.

Most Webflow shops stop at the marketing site. Most development firms don’t understand brand. When you find a team that does both, and is willing to show you the guts of how they do it, the 40-question RFP starts to look like a child’s toy.

We have become so obsessed with the “weighted score” that we’ve forgotten what we are actually buying. We aren’t buying a vendor; we are buying a future state of our business. We are buying the ability to move fast in Q3. We are buying the confidence that our brand will look as good on a customer’s phone as it does on our 32-inch studio monitors.

We are buying the certainty that when an AI crawler hits our site, it will find structured data that actually makes sense.

The Friday Afternoon Test

If you are currently sitting in front of a spreadsheet, staring at four columns that look nearly identical, I want you to do something uncomfortable.

Close the spreadsheet. Stop looking at the “Years in Business” column.

Pick up your phone and go to the “Work” page of the two agencies at the top of your list. Navigate through their latest three launches. Try to find the contact page. Try to read a blog post on a shaky Wi-Fi connection.

Then, ask yourself: “If I needed to change the hero image on this site at 4:00 PM on a Friday, would I be able to do it myself, or would I be at the mercy of a support ticket?”

The answer to that question isn’t in your rubric. It’s not in the liability insurance certificate. It’s in the architecture of the site itself.

The organizations that end up with the best results are almost always the ones that treat the vendor selection as a search for a partner, not a search for a supplier. A supplier is someone who fulfills a set of requirements you’ve written down.

A partner is someone who tells you that your requirements are wrong because they lead to a site that will be obsolete in two years. Procurement, by its very nature, penalizes the partner and rewards the supplier. It rewards the firm that says “Yes” to every poorly conceived question in the RFP, and it filters out the firm that says, “We don’t do it that way because it will hurt your SEO in the long run.”

Looking for the Merit-Based Recommendation

I’ve learned to look for the “merit-based recommendation.” This is the moment when a vendor tells you to stay on a certain platform or move to a custom code solution based on what’s best for the project, not based on what they happen to have on the bench that week.

If a firm only does Webflow, they will tell you Webflow is the answer to every problem. If they only do custom React apps, they will tell you that a simple marketing site needs a headless architecture that will cost you

to maintain.

We need to stop asking if the vendor has been around for ten years and start asking if the people who are in the room during the pitch are the same people who will be writing the CSS and configuring the CMS collections. We need to stop asking for three references in our vertical and start asking to see a screen recording of the backend of a recent build. We need to stop valuing the “defensibility” of the choice and start valuing the “usability” of the result.

The next time you find yourself adding a “tiebreak” tab to a spreadsheet, remember that you aren’t just choosing a price. You are choosing the level of frustration your marketing team will feel every day for the next three years.

You are choosing whether your brand will be a cohesive, high-performance engine or a fragmented collection of “good enough” parts.

The spreadsheet will never tell you the truth. It will only tell you what you need to hear to feel safe in the meeting.

Real quality is found in the lines of code, in the logic of the CMS, and in the transparency of the people doing the work. Everything else is just paperwork.