Quirks by Design: Privacy Thresholds

Why These "Limitations" Exist

Last updated 26 days ago

ProcureValue has several intentional design decisions that protect respondent privacy. These might seem like "quirks" at first, but they're carefully designed features.

Why Can't I Send Surveys to Just 5 People?

Minimum Requirement: 10 invitations per campaign

The Privacy Math

If you invite only 5 people and 4 respond:

  • Response rate = 80%

  • High probability you can guess who the non-responder is

  • Even anonymous responses become "semi-identifiable"

Example:

You invite: Sarah, Tom, Ahmed, Li, and Priya
4 people respond anonymously
You know exactly who didn't respond: one of them
Responses become less honest because participants fear identification

The Solution

By requiring 10 minimum invitations, we ensure:

  • Larger pool makes individual identification much harder

  • Response rates become less predictable (40%? 70%? Unknown!)

  • Participants feel safer providing honest feedback

  • Statistical validity improves

Why Can't I See Analytics With Only 3 Responses?

Minimum Requirement: 5 responses for analytics display

Small Sample Problems

With only 2-4 responses:

  • Statistical noise dominates — One extreme response skews everything

  • De-anonymization risk — Comments might be identifiable

  • False confidence — You might draw wrong conclusions from limited data

Example:

Campaign with 3 responses:

  • Response 1: PVX = 9/10 (Very positive)

  • Response 2: PVX = 2/10 (Very negative)

  • Response 3: PVX = 7/10 (Positive)

Average PVX = 6/10

But is this meaningful? No!

  • Could be one angry person + two satisfied people

  • Or one person having a bad week

  • Sample is too small to draw conclusions

The Solution

By requiring 5 minimum responses, we ensure:

  • More stable statistical measures

  • Reduced impact of outliers

  • Better protection against comment identification

  • Confidence that insights reflect patterns, not noise

Why Focus on the Organization, Not Individuals?

Design Philosophy: Tenant-level insights, not individual performance reviews

The Problem With Individual Tracking

Traditional feedback systems often:

  • Create defensive behaviors ("I need to look good")

  • Discourage honest feedback (fear of retaliation)

  • Turn into HR tools instead of improvement tools

  • Measure individuals when the system is the real problem

Example:

If Sarah in procurement gets "low scores," is it because:

  • Sarah has skills gaps? (Individual problem)

  • Sarah is understaffed? (Organizational problem)

  • Sarah's process is broken? (System problem)

  • Sarah inherited a mess? (Historical problem)

Without context, individual metrics are useless or harmful.

The ProcureValue Approach

We measure:

  • How the procurement function is perceived overall

  • What categories need improvement (process, responsiveness, expertise)

  • Which themes appear across responses (AI-identified patterns)

  • Organizational capability gaps (training needs, resource issues)

We don't track:

  • Which specific person got what feedback

  • Individual performance scores

  • Who responded vs. who didn't

  • Time-stamps that could identify respondents

Why These Thresholds?

Balancing Privacy vs. Utility

ThresholdToo LowJust RightToo HighMinimum invitations5: Easy to identify non-responders10: Hard to identify individuals50: Unrealistic for many teamsMinimum responses2: Statistically meaningless5: Stable patterns emerge20: Too hard to achieve

Research-Backed Numbers

These thresholds are based on:

  • Privacy research: 10+ individuals = "k-anonymity" protection

  • Statistical validity: 5+ samples = meaningful patterns can emerge

  • Practical reality: Most procurement teams can achieve these numbers

  • GDPR compliance: Privacy-by-design principles

Common Questions

"But I only have 8 suppliers in this category!"

Options:

  1. Broaden the category (e.g., "All IT suppliers" instead of "SaaS suppliers only")

  2. Run the campaign less frequently (annual instead of quarterly)

  3. Combine stakeholder + supplier surveys to reach 10 invitations

  4. Accept that very small segments can't be measured anonymously

Remember: Privacy protection is more important than measuring every micro-segment.

"Can I override the thresholds?"

No. These are hard limits to protect respondent privacy.

Even as an administrator, you cannot:

  • View campaigns with <10 invitations

  • See analytics with <5 responses

  • Identify individual respondents (ever, for any reason)

"What if people just don't respond?"

This is valuable data!

Low response rates tell you:

  • Survey fatigue (too frequent campaigns)

  • Lack of trust in anonymity claims

  • No perceived benefit to participating

  • Poor timing (busy season)

Improvement strategies:

  • Communicate how feedback drives change

  • Share what actions you took from previous surveys

  • Keep surveys short (ProcureValue surveys are ~5 minutes)

  • Choose timing carefully (avoid month-end, fiscal year close)

"Can I see who responded?"

Absolutely not. This would defeat the entire purpose.

ProcureValue is designed so that:

  • No one (including admins) can see who responded

  • Responses cannot be correlated with invitations

  • Even with database access, linkage is impossible

If you need individual feedback, use a different tool (1-on-1s, performance reviews, named surveys).

The Bottom Line

These "quirks" are intentional features that:

  • ✅ Protect respondent privacy

  • ✅ Ensure statistical validity

  • ✅ Focus on organizational improvement

  • ✅ Build trust in anonymity claims

  • ✅ Comply with GDPR/privacy regulations

If you could override these, the system would lose its value.

Honest feedback requires genuine anonymity. These thresholds ensure it.

Still have questions? Contact your ProcureValue administrator or email support@procurevalue.com.