Line up what each part does against what it costs, and stop paying for value nobody gets.
Every product, service or process is a bundle of parts, and every part costs something. Value engineering asks a simple question of each one: what job does this do, and is that job worth what we pay for it? In this view, value is how well something performs the functions people need, divided by what it costs.
You describe each function in two words, a verb and a noun, like hold coffee or stop spills. You find out which functions matter most to the people who use it, and you put a cost on every part. Lining the two up shows where money buys little and where a small investment could buy a lot. The aim isn't simply to cut. It's to deliver what people value, for less.
Strictly, value analysis improves something that already exists and value engineering applies the same thinking at the design stage. In practice, most people use the two names interchangeably.
Use it when costs are rising and you can't just raise prices, when you suspect some features survive out of habit, or when you need to free up time or money without hurting what customers care about.
Function–cost matrix for a café's takeaway cup. The 8¢ sticker does one minor job, so look there first.
What exactly are you analyzing, what is it made of, and who uses or pays for it?
What jobs does it do, and which one is the reason it exists?
Which functions matter most to the people who use it, and which barely register?
What does each part cost, counting materials, labor, waste and handling?
Where are you paying a lot for a function people value little, or too little for one they love?
What else could perform this function, and what would that cost?
Most cost-cutting starts with the price list. Value engineering starts with the job each part does, and that change of question reveals options a price list never will.
Ask about “a sticker” and the obvious answer is a cheaper sticker. Ask about “show brand” and dozens of answers appear: a printed sleeve, a stamp, a cup color, a memorable barista. Naming the job instead of the part opens up better options.
Costs are tracked part by part, while value is felt by customers, so the two rarely meet. A function–cost matrix puts them on one page, and the mismatches become hard to miss.
Insiders tend to overrate the features they worked hardest on. The people who use the thing often rank its functions quite differently, and the gap between the two lists is where savings, or smart extra spending, hide.
Because you start from the functions people need, you cut the cost of delivering them, not the functions themselves. That protects you from savings that quietly drive customers away.
Pick one product, service or process, and name the people who use or pay for it. List its parts or steps. Keep the scope small enough to cost properly: one service, not the whole shop.
Describe each job it does as a verb plus a noun: hold coffee, explain work. Mark the one basic function it exists for. The rest are supporting functions.
Ask a handful of real users to rank the functions, or watch what they choose. Insiders and users often rank things differently, and the users' ranking is the one that counts.
Put a cost on each part or step: materials, labor minutes, waste and handling. Use real figures where you have them and mark the estimates.
Draw a grid with parts down the side and functions across the top. Show which part delivers which function, and at what cost. Circle parts that cost a lot but serve low-value functions.
For each circled part, ask: what else could do this job, and what would that cost? Drop, merge, simplify or swap. Then look for valued functions worth spending more on.
A list of parts or steps, real cost figures, a few customers to ask, and a spreadsheet or a grid on paper. With AI, the prompt below builds the matrix with you.
Subject: the $75 standard tune-up at a two-mechanic bike shop. Goal: same-day service without cutting quality. Constraint: no new staff or equipment.
A standard tune-up: safety check, brake and gear adjustment, wheel truing, a full wash and polish, and a written report. About 80 minutes of labor.
Basic function: make bike safe. Supporting functions: shift smoothly, roll true, look clean, explain work.
Twenty customers ranked the functions. Safety and smooth shifting came first, a clean frame came last, and many said speed mattered almost as much as the work.
Labor in minutes: safety check 10, brakes 15, gears 20, wheels 10, wash and polish 20, report 5.
The wash and polish used a quarter of the labor for the function customers valued least.
Make the full wash a $15 add-on and give every bike a quick wipe instead. The 15 minutes saved per bike open two same-day slots a day.
Tune-up labor drops from 80 to 65 minutes, with no change to the checks customers care about most. Riders who want the full wash can still buy it, and two same-day slots open up every afternoon.
Value engineering involves careful bookkeeping: naming functions, building a matrix and comparing alternatives. AI handles that structure well and suggests functions and alternatives you might miss. It can't know your real costs or what your customers value. Bring the numbers and the customer's voice; let AI do the organizing.
Turn a parts list into clear verb-noun functions.
List the functions of each step in our bike tune-up as a verb plus a noun. Mark the basic one.
Line up parts, functions and costs in one table.
Build a function-cost matrix from these parts and costs. Show each part's share of the total.
Find other ways to deliver one costly function.
Give me 8 cheaper ways to 'show brand' on a takeaway cup, each with a rough trade-off.
Look for places where spending a little more raises value a lot.
Which high-value function is underfunded here? Suggest 3 ways a small extra spend could improve it.
Describe the item and what you know about its costs, pick a mode, then copy the prompt into Claude, ChatGPT, Gemini or Copilot. In coach mode the AI works through the value questions with you one at a time and builds the function–cost matrix as you go.
The “Open in” links put the prompt in the web address. If your challenge includes anything sensitive, copy and paste instead.
After the session, click one to copy it, then paste it into the same chat.