FlowStatus

Benchmark study · Wave 1 · September 2026

How design teams work: focus is the leading indicator of project health

Findings from the first wave of the Enterprise Design Telemetry Benchmark: 43 design projects in six organisations, measured passively through Figma activity logs.

Published:
25 September 2026
By:
MoreThought, makers of FlowStatus
Sample:
43 projects · 6 organisations

Summary

43
design projects analysed
6
participating organisations
3.1 vs 2.1
outcome score, high vs low focus (out of 5)
0.97–0.98
week-three correlation with full project

The findings

Across 43 design projects in six organisations, focus emerged as the strongest predictor of project success among the factors we examined. Focus, measured by whether designers sustained attention on individual files or fragmented their effort across many, explained more variation in outcomes than team size, seniority, tooling or process.

Projects with low file-switching averaged 3.1 out of 5 on business and customer outcomes. Projects with high file-switching averaged 2.1. That gap, a full point on a five-point scale, held regardless of team size.

And it is visible early. The same behavioural signature that predicts final outcomes is already present in the first three weeks of a project. Organisations do not need to wait for a project to finish to know whether it is in trouble. Week-three metrics for activity, consistency and file-switching correlate at 0.97–0.98 with the full project, so the early read is representative.

Why this matters

Design leaders currently have almost no leading indicators. Outcome data (whether a project shipped well, whether it moved the metrics it was supposed to) arrives too late to act on. Everything else available today, such as velocity, headcount and ticket counts, measures activity, not coherence.

Focus is different. It is a measurable proxy for the cognitive conditions under which good design work actually happens: sustained, undivided attention on a problem, rather than shallow, repeated context-switching across many. This is consistent with research on flow states and attention residue: when people switch tasks, part of their cognitive capacity stays attached to the task they left, degrading performance on the one they moved to. File-switching in Figma is an unusually clean behavioural trace of exactly this pattern, captured automatically and at scale.

This converts an abstract management concern about team focus into a real-time number, available three weeks into a twelve-week delivery cycle instead of at a retrospective.

The evidence

Two modes of working emerged clearly when projects were grouped by behaviour.

Table 1. Project outcomes by focus group
MeasureHigh focusMediumLow focus
Projects151117
Average outcome score (1–5)3.12.72.1
File-switchingVery lowModerateHigh
Files worked on by multiple people36%29%23%
Groups are relative to the 43-project sample, not industry norms. Outcome scores are business and customer outcomes, self-reported by participating teams.
012345High focus15 projects3.1Medium11 projects2.7Low focus17 projects2.1Average outcome score (1–5)
Figure 1. Average business and customer outcome score by focus group, 43 projects. Source: Design Telemetry Benchmark 2026, Wave 1.

High-focus projects were not simply smaller or better resourced; team size did not explain the gap. They were distinguished by fewer people jumping between files and more files receiving sustained, shared attention.

How we tested alternative explanations

We ran four separate models to see what best predicts project outcomes. File-switching consistently outperformed the obvious alternatives.

Table 2. Competing models of project outcomes
What we testedMeasures usedPseudo-R²Finding
File-switching + team sizeSwitching rate, number of designers0.21Switching dominates; adding people does not compensate
Switching aloneSwitching rate only0.15Strong predictor even without controls
Duration + momentumProject length, consistency of work0.11Both help, but jointly weaker than switching alone
Team compositionEditors versus viewers0.11Viewer-heavy teams score lower
Pseudo-R² is McFadden's pseudo-R², a measure of model fit. Higher means the factors explain more of the variation in outcomes. McFadden pseudo-R² values run much lower than ordinary R²; McFadden described 0.2–0.4 as an excellent fit.

The effect persists after accounting for the obvious confounders. Team size, project duration and the proportion of active editors were all tested as alternative explanations. File-switching remained the dominant factor even when these were controlled for.

Separately, early collaboration on shared files accounted for a meaningful share of outcome variation, a consequential effect for a single early-stage signal.

The week-three signal

The signal appears early enough to act on. Early-window metrics (the first 21 days) for activity, consistency and file-switching correlated at 0.97–0.98 with full-project metrics. A three-week read is, in practical terms, the whole story.

Table 3. How well does week three predict the full project?
Week-three signalCorrelation with final project signal
Total activity0.98
Work consistency (momentum)0.98
File-switching0.97
A near-perfect correlation means the first three weeks are representative. There is no hidden pattern that emerges later to reverse the early read.

How focus is measured

Focus (switch rate) is the share of a project's Figma activity where a designer's very next logged action is on a different file. Values range from 0 to 1: higher means more file-hopping.

High and low groups are relative to the 43-project sample, not industry norms. Organisations with heavy viewer traffic may show a lower number for reasons unrelated to focus. For how FlowStatus builds on this signal in the product, see the methodology page.

Study design

Table 4. Study design at a glance
ElementWave 1 approach
Sample43 design projects across six organisations
Data sourcePassive Figma activity-log telemetry: actor, timestamp, action type and file, with no surveys, plugins or access to design contents
Outcome measureBusiness and customer outcome score on a 1–5 scale, self-reported by participating teams
Focus measureSwitch rate: share of actions where the next action by the same designer is on a different file
Other variablesTeam size, project duration, momentum (consistency of work), editor versus viewer composition, shared-file collaboration
GroupingHigh, medium and low focus, relative to the sample distribution
Model evaluationMcFadden's pseudo-R² across four competing models
Early-window testCorrelation between first-21-day metrics and full-project metrics

Organisations operated different design models, including centralised pools, embedded squads and design system teams. For how to get the same activity data out of your own Figma organisation, see the Figma activity audit guide.

What design leaders should do

1. Run a focus check at week three

If designers are bouncing across many files with little overlap, intervene on scope and stakeholder alignment before adding headcount. Adding people to an unfocused project compounds the problem.

2. Stop equating visibility with contribution

A large roster of Figma viewers is not a sign of health. Track who is actually building, not who is watching.

3. Don't apply one collaboration template everywhere

Different operating models (centralised pools, embedded squads, design system teams) reach good outcomes through different patterns. The focus signal travels across all of them; specific collaboration norms do not.

What we're confident about, and what we're not

This is the first wave of an ongoing benchmark. The headline finding, that focus predicts outcomes and predicts them early, is consistent, holds after controlling for team size, and is corroborated by a separate early-window analysis. We're confident in it.

What's next

The methodology can also interpret the impact of momentum (how consistently a project is worked on), team size, collaboration and participation, iteration rounds and speed. Early signals showed some effects on outcomes, but the sample was too small in this round to report them with confidence, so a second wave is planned.

Wave 2 opens in October 2026. It will expand the sample and test these signals against independent performance data. If you lead design teams and want to participate, email hello@morethought.co or use the button at the foot of this page.

Download the original report (PDF, 8 pages)

Frequently asked questions

What is the strongest predictor of design project success?

In the 2026 Design Telemetry Benchmark (43 projects, six organisations), focus was the strongest predictor of project outcomes among the factors examined. Focus was measured as file-switching in Figma: projects where designers sustained attention on individual files outperformed projects where effort was fragmented across many files. It explained more variation in outcomes than team size, seniority, tooling or process.

How much better do high-focus design projects perform?

High-focus projects (low file-switching) averaged 3.1 out of 5 on business and customer outcomes. Low-focus projects (high file-switching) averaged 2.1 out of 5. Medium-focus projects averaged 2.7. The one-point gap held regardless of team size.

How early can you tell whether a design project is in trouble?

By week three. Activity, work consistency (momentum) and file-switching measured in the first 21 days of a project correlated at 0.97 to 0.98 with the same metrics measured over the full project. The behavioural signature that predicts final outcomes is already present in the first three weeks.

Does adding designers fix an unfocused design project?

The benchmark suggests not. In the model combining file-switching and team size (McFadden pseudo-R² 0.21), switching dominated and adding people did not compensate for low focus. The recommendation is to intervene on scope and stakeholder alignment before adding headcount, because adding people to an unfocused project compounds the problem.

What is file-switching or switch rate in Figma?

Switch rate is the share of a project's Figma activity where a designer's very next logged action is on a different file. It ranges from 0 to 1; higher values mean more file-hopping and lower focus. It is derived passively from Figma activity logs, with no surveys or plugins.

Does the benchmark prove that focus causes better outcomes?

No. The benchmark shows a consistent, early and robust association, not strict causation. Chaotic projects could generate switching rather than the reverse, although the early-prediction result makes pure reverse causation less likely. Outcomes in Wave 1 were self-reported by participating teams.

Can my organisation take part in the next benchmark wave?

Yes. Wave 2 opens in October 2026 and will expand the sample and test the signals against independent performance data. Design leaders can register interest by emailing hello@morethought.co or using the Talk to us form on flowstatus.io.

Cite this study

Suggested citation

MoreThought (2026). How Design Teams Work: Focus is the leading indicator of project health. Enterprise Design Telemetry Benchmark, Wave 1. FlowStatus. https://flowstatus.io/telemetry-benchmark

About MoreThought

MoreThought provides operational telemetry and diagnostic intelligence for enterprise product and design organisations, and builds FlowStatus. We help executive leaders replace subjective advocacy, manual spreadsheets and disruptive workshops with objective activity data. By analysing passive workflow telemetry from the tools teams use every day, we expose where delivery stalls, quantify true team capacity and isolate cross-functional friction without interrupting ongoing work.

Run the week-three focus check on your own projects

FlowStatus computes the same signals from your Figma activity logs, continuously, for every team and project. Book a walkthrough, or register to take part in Wave 2 of the benchmark.

Home · How it works · Design Telemetry Benchmark · Methodology · Figma activity audit guide · Log in · © 2026 FlowStatus by MoreThought