This section outlines the conceptual framework to give an understanding of the performance analysis of startups accelerated by Google and Microsoft. Built to align with the ANOVA (Analysis of Variance) based methodology adopted in this dissertation, the framework maps the key components, defines the performance indicators used in the study, and explains the logic behind time-based segmentation and the reasoning behind its group comparisons.
This post represents a series of articles related to a research and dissertation called “Are corporate accelerators springboards for startups: a performance analysis of the Microsoft’s and Google’s accelerated.
PURPOSE OF a Conceptual Framework for Startup Accelerators
The framework serves as a foundation to evaluate whether participation in corporate accelerator programs results in observable, measurable differences in startup performance (which variables are explained later). It is structured as follows:
- Results in certain periods or stages divided in five standard “periods of time”.
- Comparison between two distinct groups: Google vs. Microsoft.
- Since ANOVA is the main analysis, to figure out whether performance differences are statistically significant.
STRUCTURE OF THE FRAMEWORK
This conceptual model is built around three core elements.
Grouping Criteria
Startups are categorised by their affiliation to a corporate accelerator:
- Google (617 startups).
- Microsoft (238 startups).
Time Dimension
Startup performance is segmented into five periods of time. In parentheses, how the label is displayed in the graphs:
- Before Acceleration (“Funds raised at the acceleration date (M)”).
- At Acceleration Date (“Amount of funding_ Accelerator at the Acceleration date”).
- Year 1 (“Amount of Funding (year 1)”).
- Year 2 (“Amount of Funding (year 2)”).
- Year 3 (“Amount of Funding (year 3)”).
Observed Outcomes
These are the performance signals measured and tested across each period. To increase the perception or understanding of each outcome, they are divided into sections. Those are:
- Funding (in USD millions).
- IPO (Initial Public Offering) occurrence.
- Acquisitions (whether the startup was acquired).
- Acquisitions Made (whether the startup acquired another company).
- Closure status (whether if startup was closed or remained active).
Each of these indicators are directly retrieved from the Crunchbase database and reported in either monetary values (continuous) or binary (0/1) format, depending on the variable.
TIME Segmentation LOGIC
Each period of time represents a unique phase in the startup lifecycle in their participation of the accelerator program:
- Before Acceleration: Indicates prior investment, maturity, and external support, before the arrival of Google or Microsoft.
- At Acceleration Date: Displays formal support by the corporate accelerator program (initial financial injection, mentoring push or milestone funding).
- Years 1–3: Allow measurement of continuous funding, survival and evolution.
By repeating the same outcome measures across all periods, the framework enables a structured comparison of average startup performance within and across accelerator groups.
VIsual Representation
Below is located a simple table that reinforces and fits what was told earlier in the conceptual framework for startup accelerators:
| Variable | Component | Description |
|---|---|---|
| Grouping | Accelerator Affiliation | Google vs. Microsoft |
| Time Dimension | Time Periods | Before Acceleration, At Acceleration, Year 1, Year 2, Year 3 |
| Outcomes Measured | Performance Metrics | Funding (USD), IPOs, Acquisitions, Closures |
| Comparison Objective | Mean Comparison | Do average values differ between Google and Microsoft over time? Do they represent a measurable performance advantage? |
| Statistical Method | One-way ANOVA | Used to determine if group differences in means are statistically significant |
WHY ANOVA FITS THIS FRAMEWORK
The model of the conceptual framework for startup accelerators is structured to fit ANOVA because it tests whether the mean of a given outcome variable (e.g., funding or IPO frequency) differs significantly between two or more groups. It is appropriate here because group means between Google vs. Microsoft are compared and fixed time intervals are established.
CONCLUSION
This conceptual framework for startup accelerators links the strategic evaluation of accelerator programs to empirical and objective testing through the lens of ANOVA and miscellaneous testing. With time-based outcomes (periods of time) and grouping (startups by accelerator), it allows for a neutral, replicable comparison of Google vs. Microsoft accelerators without relying on causal prediction. In plain English, ANOVA does not aim to relate or prove causation or cause-and-effect relationship, but to compare averages or means between its groups. It also prepares the understanding pillars for Chapter 4.
All variables referenced are further explained in Section 3.3 (Data Collection Methods).
