Most small businesses compare vendors by lining up three quotes and picking the lowest number. That works right up until the cheap vendor misses a deadline, ships a batch you have to re-check by hand, or stops answering email for nine days. The price on the page was real. The cost was not.
This is an information problem, not a negotiation problem. You are making a prediction about future behavior using a document that only describes price. The fix is to measure the things that actually predict behavior — before you sign, and every month after.
You can build the whole system in a blank spreadsheet this afternoon.
Why the Cheapest Quote Is Rarely the Cheapest Vendor
A quote prices the deliverable. It does not price your time.
Every vendor relationship carries an internal cost that never shows up on an invoice: chasing status updates, re-explaining a spec you already sent, inspecting work that arrived wrong, rescheduling the tasks that sat behind it, and apologizing to your own customer. That cost is paid in your hours.
So stop comparing quotes and start comparing landed cost:
Landed cost = quoted price + (rework hours × your hourly rate) + (admin and chasing hours × your rate) + cost of delay.
Pick an hourly rate for your own time and use the same one for every vendor. It does not need to be perfect. It needs to be consistent, so the comparison is honest.
For a first-time vendor you have no history, so you estimate. For a vendor you already use, you should not estimate — you should look it up. That is what the log in section four is for.
The Five Factors That Predict Vendor Performance
Score every vendor on the same five things. Adding a sixth rarely changes the ranking; dropping one usually does.
- Delivery: did the work arrive on the agreed date, and were you warned in advance when it would not?
- Quality: what share of the work was accepted without rework on the first pass?
- Responsiveness: how long from your message to a real answer — not an auto-reply?
- Cost: landed cost, not quoted price, relative to the alternatives.
- Documented issues: how many problems were logged in the period, and how were they resolved?
Anchor the 1–5 scale in writing
A score means nothing unless the scale is defined before you use it. Write the anchors once and reuse them. For delivery, for example:
- 5 — on or before the agreed date, every order in the period.
- 4 — late once, under 48 hours, with advance notice.
- 3 — late occasionally with notice, or once late without notice.
- 2 — late repeatedly, notice inconsistent.
- 1 — late without notice, or a missed delivery you found out about from your customer.
Write the same five lines for quality, responsiveness, cost and issues. The anchors are the actual work. Once they exist, scoring takes minutes and two different people produce nearly the same number.
How to Compare Vendors With a Weighted Scorecard
Not every factor matters equally. If you resell a physical product, delivery is close to everything. If you buy design work, quality and responsiveness outrank a $200 price gap.
Build it like this in a blank sheet:
- Row 1: the five factor names across columns B to F.
- Row 2: your weights, as whole numbers that add to 100. Set these before you look at any quotes.
- Rows 4 and down: one vendor per row, with a 1–5 score in each factor column.
- Column G: the weighted total, using SUMPRODUCT of the score range and the weight range, divided by 100.
That gives you a single number between 1 and 5 for each vendor, built from criteria you committed to in advance. Setting the weights first is what stops you from quietly re-weighting until your favorite wins.
When you compare vendors this way, you also get something a quote never gives you: a written record of why you chose. Six months later, when someone asks why you are paying more than the lowest bidder, the answer is on one screen.
A Worked Example: Three Vendors for a Recurring Print Job
Weights: delivery 30, quality 25, cost 20, responsiveness 15, documented issues 10. Your internal rate is $65 an hour. Quantities are identical across all three quotes.
- Vendor A — quote $1,800. Delivery 2, quality 3, cost 5, responsiveness 2, issues 3. Weighted score 2.95.
- Vendor B — quote $2,150. Delivery 4, quality 4, cost 4, responsiveness 5, issues 4. Weighted score 4.15.
- Vendor C — quote $2,400. Delivery 5, quality 5, cost 3, responsiveness 3, issues 5. Weighted score 4.30.
Now add your hours. Vendor A typically costs you five hours of rework and three hours of chasing per run: 8 × $65 = $520, so landed cost is $2,320. Vendor B costs about one hour: $2,215. Vendor C costs about half an hour: $2,433.
The lowest quote is the second most expensive option once your time is counted. That is the whole point of the exercise, and you cannot see it from a quote comparison.
Between B and C the scores are 0.15 apart — inside the noise of any 1–5 scale. So the tiebreaker is landed cost, and Vendor B wins.
The decision rule
Write this down and follow it:
- If one vendor leads by 0.5 or more on weighted score, choose that vendor even if it is not the cheapest quote.
- If the top two are within 0.5, choose the lower landed cost.
- If any vendor scores 1 or 2 on delivery or quality, it is out — regardless of total score or price.
That last line is the one that saves you. A cheap vendor can pile up enough points on cost to look acceptable in aggregate while failing at the only two things you actually bought.
Turn the Scorecard Into a Running Record
A one-time comparison decays fast. Vendors change staff, take on bigger clients, and quietly deprioritize you. Selection and monitoring are the same system.
Add a second sheet with one row per order: date, PO or job reference, promised date, actual date, first-pass accepted (yes/no), hours of rework, hours spent chasing, response time in hours, and a free-text issue note.
Log it the day it happens, not at quarter end. Memory grades on the last bad experience; a log grades on the period.
Then re-score quarterly from the log, never from recollection. Two thresholds are enough:
- Below 3.5 — a documented conversation with the vendor, specific examples, a date to review.
- Below 3.0 for two consecutive quarters — start sourcing a backup. Do not wait for the failure that forces it.
Share the numbers with the vendor. Most suppliers respond well to being measured against anchors they can see, and a scoring conversation is far easier than a termination one. We cover more operational systems like this on the Cursiqa blog.
Where a Spreadsheet Stops Being Enough
Be honest about the limits before you hit them.
A spreadsheet handles this well up to roughly fifteen or twenty active vendors, with one or two people entering data. Past that, four things break: nobody knows which file is current, entries stop being logged the day they happen, you cannot restrict who edits historical scores, and renewal dates get missed because nothing sends a reminder.
Move to dedicated procurement or vendor management software when you need an audit trail you can defend, purchase orders matched against your accounting system, contract renewal alerts, or role-based access because more than two people touch the data. If you operate in a regulated supply chain, that threshold arrives earlier.
The biggest failure mode is simpler than any of those: the log goes stale. A scorecard built on three weeks of data and nine months of guessing is worse than no scorecard, because it looks objective. If you are not going to log issues, do not pretend to score them — just track delivery dates and landed cost, and be right about two things instead of vague about five.
If you would rather start from a built sheet than a blank one, the Vendor Selection & Performance Scorecard ships the weighted comparison, the anchored scales and the running log already wired together, and the rest of the operational workbooks are in the Cursiqa shop. Either way, run your next three quotes through the five factors before you sign anything.