Marcus Andrey Study
Ler em português Library

Design Systems · For leads · 5 min read

Showing the value of a design system

For leads: how to show what a design system is worth with numbers you can defend. What the one public timed study found and what it didn't, how to measure speed, quality and consistency in your own team, and how to report it without inventing ROI.

"Design systems save time" is true and almost useless in a budget meeting. Everyone has heard it, and nobody can check it. What works is smaller and stronger: a number from your own team, measured in a way a sceptical colleague could repeat, next to an honest line about what the system costs. This lesson is how to get that number, and the other kinds of value that don't fit in a sum.

What you'll be able to do after reading:

  • Name the three kinds of value a system brings
  • Read a published study for what it does and doesn't prove
  • Measure speed, quality and consistency in your own team
  • Report value with its costs, as a worked sum and a story

1. Three kinds of value

  • Speed. Screens and features get built faster, because the parts exist.
  • Quality. Accessibility, states and edge cases are solved once, in the component, and every product gets the fix.
  • Coherence. Products look and behave like one company, and people don't have to relearn each one.

Speed is the easiest to count, so it gets all the attention. Quality is often the strongest argument, because it has a cost everyone recognises: the same bug, fixed three times in three products.


2. What the published study says

In 2019 Figma published a study with its own design team: each designer did two matched tasks, one with a design system and one with old design files to reference, and they completed the objective 34% faster with the design system (Figma, Measuring the value of design systems, 2019).

Read the whole article before you quote it. The author is clear about its limits:

  • The system was directly applicable to the task, up to date and relevant, so they frame the result as the maximum time savings you'd see in practice; real teams are "probably somewhere in between".
  • The participants were Figma's own designers, and the article doesn't give their number or a statistical test.
  • It measured design time only, not engineering, review or maintenance.

It's a good method and a ceiling, not an expected return. Use the method; measure your own number.


3. Measure your own speed

Worked example: the numbers are invented, the method is not
LineWhere the number comes fromExample
Screens built or changed per quarterYour tickets or design files, counted40
Hours per screen without HarbourA timed task on old files, the median of your designers7.5 h
Hours per screen with HarbourThe same people, a matched task, with the library5.0 h
Design hours saved40 × (7.5 − 5.0)100 h
Minus time spent on the system by product teamsContributions, migrations, office hours, from timesheets or estimates− 30 h
Net design hours back, per quarterBefore the system team's own cost, which you show next to it70 h
A value sum a finance partner can check line by line. Every input is something you can count or time in your own team, and the system's cost is shown next to it, not hidden. Hours are only one line of the story: quality and consistency go beside it.Example: Marcus Andrey, CC BY 4.0. Timed-task method after Figma's 2019 study, "Measuring the value of design systems".

You can repeat Figma's method on a small scale:

  1. Pick two matched tasks from your real backlog: two screens of similar size and difficulty.
  2. Each designer does one with the library and one without, in alternating order, so each person is their own comparison.
  3. Time them, and let each designer decide when they're done, so quality stays comparable.
  4. Use the median, not the average, with a handful of people.
  5. Multiply by real volume: how many screens your teams actually build per quarter.

Then subtract what the system costs product teams: contributions, migrations, the time spent asking questions. Show the system team's own cost next to the result, not hidden. A finance partner trusts a sum with a minus sign in it.


4. Measure quality and consistency

These are counts, so they are easy to defend:

Measure How to get it Harbour example
Repeated bugs Accessibility or UI bugs fixed once in a component instead of once per product The focus ring fixed in Button reached both products in one release
Distinct versions Rerun the inventory a year later (see Run an interface inventory) Blues: 14 to 3. Buttons: 9 to 1 component
Adoption Share of screens or imports that use the system (see Measuring adoption) Booking app 90%, console 60%
Measured contrast Pairs checked and passing, per brand and mode Every Harbour text pair at or above 4.5:1
Time to brand How long a new brand took to launch Skiff: new values, no forked components

The Harbour numbers are the course's fictional example. Yours will be messier, and that's fine: a messy real number beats a clean invented one.

Sparkbox's 2022 survey found that only 16% of respondents tracked metrics for their system at all (Sparkbox, Design Systems Survey 2022). Measuring anything puts you ahead.


5. Tell it as a story with numbers

Numbers convince the budget owner; stories convince everyone else. Pair them:

  • One sum, like the figure, with its sources and its costs.
  • One before and after: the same screen in last year's product and this year's.
  • One quote from a product designer or engineer about what got easier.
  • One thing that's still hard, and what's on the roadmap about it (see A roadmap for the system).

Report on a rhythm, every quarter or half year, in the same format, so the trend matters more than any single number.


6. Common mistakes

  • Quoting "34% faster" as what your team will get.
  • Industry ROI figures with no traceable study behind them.
  • Counting only design time, and ignoring engineering and maintenance.
  • Hiding the system's cost, so the first question from finance breaks the case.
  • Measuring once, at launch, and never again.

7. Try it (no tools needed)

  1. Pick the one measure from section 4 you could count this week.
  2. Write where the number would come from, exactly: which tool, which filter, which person.
  3. Write the sentence you'd say about it in a budget meeting, including one limit.

Reflection questions: Which kind of value would matter most to the person who funds your system? What would you need to measure to show the system isn't worth it, and are you willing to?


Checklist

Learn more

Books, courses and videos to go deeper

Book links go to Amazon. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.

Measuring adoption · Funding and staffing a design system team · A roadmap for the system · Run an interface inventory · Defining success · What a design system solves · Design system ROI estimate · Design system metrics

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close