Skip to main content

​
📊 Summary

High-quality data is the foundation of Backerville’s collective intelligence system. Our rigorous collection process ensures reliable inputs for the algorithm.
Data Collection Flow
The foundation of our algorithm is high-quality, representative data from startup employees. This page explains how we collect, validate, and process that data.

​
✅ Eligibility Requirements

Tenure

Minimum 12 months at current company

Verification

Corporate email and/or LinkedIn verification

Participation

One vote set per employee per cycle
Not all employees can participate in the voting process. We enforce specific eligibility criteria to ensure quality inputs:
Only employees with ≥12 months tenure can submit bullishness ratings. This requirement serves two important purposes:
  1. Ensures participants have meaningful context about their company
  2. Aligns with typical 1-year vesting cliffs, so participants likely have equity at stake
We use a multi-layer verification approach:
  • Corporate email verification with OAuth or confirmation code
  • Optional LinkedIn profile confirmation
  • Employment status check through company directory (when available)
  • Periodic re-verification
To maintain data integrity:
  • Each employee can submit only one set of votes per voting cycle
  • Votes cannot be changed after submission within the same cycle
  • Consistent voting patterns are monitored across cycles

​
🔄 Two Key Input Types

  • Self-Bullishness Rating
  • Peer Admiration
Sample Distribution of 85 Employee Bullishness Ratings
How bullish are you on your company's future success?

1-20:   Major concerns about viability
21-40:  Significant challenges ahead
41-60:  Moderate outlook with mixed signals
61-80:  Positive outlook with some reservations
81-100: Extremely confident in future success

Consider factors like:
- Product-market fit
- Competitive landscape
- Leadership effectiveness
- Financial runway
- Overall market conditions

​
🔍 Data Validation

Completeness Check

Ensuring all required fields are filled properly

Range Verification

Confirming ratings fall within allowed bounds

Uniqueness Check

Preventing duplicate submissions or selections

Theme Alignment

Validating that companies match the investment theme of the village
Before data enters the algorithm, it undergoes a multi-stage validation process to ensure quality and consistency.

​
⚖️ Weighting Employee Votes

ω(e)=1+0.2×min⁡(years_beyond_minimum,4)+seniority_factor\omega(e) = 1 + 0.2 \times \min(years\_beyond\_minimum, 4) + seniority\_factor
Employee Weight Calculation
Where:
  • years_beyond_minimum = max(0, tenure_years - 1)
  • seniority_factor ranges from 0 (entry level) to 0.5 (executive)
Not all employee votes are weighted equally. We apply a per-employee weight ω(e)\omega(e) based on factors that might indicate deeper insight:
Employees with longer tenure receive progressively higher weights:
  • 1 year (minimum): Base weight
  • 2 years: +0.2 weight
  • 3 years: +0.4 weight
  • 4 years: +0.6 weight
  • 5+ years: +0.8 weight (maximum tenure bonus)
Position in the company provides additional weight:
  • Entry level: No additional weight
  • Team lead/manager: +0.1 weight
  • Director/Senior Manager: +0.3 weight
  • VP/Executive: +0.5 weight
In some implementations, we may incorporate historical accuracy:
  • Based on how well an employee’s past bullishness ratings correlated with actual outcomes
  • Requires multiple cycles of data
  • Applied as a multiplier to the base weight

​
📅 Data Collection Timeframes

Our data collection follows a structured quarterly schedule:
1

Notification Phase

All eligible participants receive notifications 3 days before the collection window opens
2

Collection Window

A 2-week period when all voting occurs, with reminders sent at the beginning, middle, and 2 days before closing
3

Processing Period

Following collection, data is validated, weighted, and prepared for the algorithm
4

Results & Rebalancing

Algorithm results are finalized and rebalancing recommendations are implemented

​
🔒 Privacy & Anonymity

Anonymized Results

Individual votes never attributed to specific people

Aggregated Data

Only aggregate scores used in the algorithm

Secure Storage

Encrypted storage with strict access controls
To ensure honest feedback, we implement comprehensive privacy protections:
The privacy of individual voting data is critical to the integrity of the system. Without anonymity, employees might be reluctant to provide honest assessments, especially if they have concerns about their own company.

​
Minimum Data Thresholds

For a company to be included in the ranking calculation, we enforce minimum data requirements:

​
Full Inclusion Criteria

  • At least 3 eligible employees submitting bullishness ratings
  • At least 2 mentions from other companies in admiration lists

​
Partial Inclusion Criteria

  • 1-2 employees submitting bullishness ratings
  • At least 1 mention from another company
  • Flagged with limited confidence indicator

​
Below Threshold Handling

  • May appear in village portfolios at reduced weights
  • Marked as having limited data support
  • Receive higher uncertainty scores in confidence intervals
This rigorous data collection process ensures our algorithm has high-quality inputs to generate meaningful, manipulation-resistant rankings.