Startup Success Prediction: What Actually Drives an Acquisition

3.3x

Companies that raised a Series D round were 3.3x more likely to be acquired

~5x

Startups recognized on a top-500 list were about 5x more likely to be acquired

0.65

Seven measurable factors together explain about two-thirds of the differences in funding raised

1.4x

Startups that raised angel funding were about 1.4x more likely to close than to be acquired

What the data actually shows

Three findings that translate directly into business decisions.

Bar chart showing acquisition odds for angel/VC funding versus rounds A through D, with round D at 3.3x — Later rounds, better odds

Odds of acquisition climb through round A (2.2x) to round D (3.3x) — while angel funding leaned the other way and VC was inconclusive.

Pie chart showing a 5-to-1 ratio of acquisitions among top-500-recognized startups — Top-500 recognition matters

Startups on a top-500 list were about 5x more likely to be acquired.

Combo chart showing total and average funding raised by state, led by California and Massachusetts — Location moves the number

California leads on total funding raised; Massachusetts and California startups raised more on average than New York.

The Business Question

Two questions drove this study: what predicts whether a startup gets acquired instead of closing, and — since funding turned out to matter — what predicts how much money a startup raises in the first place. I also tested whether industry and location move the odds either way.

The Approach

I audited all 47 variables (factors) in a public dataset for range errors, contradictions, and undocumented fields, then matched each business question to the right statistical test: odds ratios for funding-round effects, chi-square for industry and state effects, t-tests and ANOVA for funding differences, and a 7-factor regression model to explain total funding raised. The seven factors: company age at its last funding round, how many investors typically joined each round, total funding rounds completed, number of strong industry relationships, top-500 recognition, angel funding, and VC backing.

Vertical flow diagram of the audit process: raw dataset, variable audit, cleaning decisions, statistical tests, and final model — Every number in this study traces back to a reproducible sheet in the workbook.

Role: Sole analyst — data audit, cleaning strategy, statistical testing, and regression modeling
Team: Solo
Timeline: Sep 2024

How Much More Money Can You Actually Raise?

Seven measurable factors

Once a startup is raising money, the amount it ends up with isn't random — a regression model on this dataset shows which factors go together with higher funding, and by how much.

The strongest signals were angel funding (much less raised), top-500 recognition (about double), and each extra funding round (+52%).

The 7 Factors Behind Funding

Angel funding: 85% lessAngel funding
Top-500 recognition: About 2× moreTop-500 recognition
Each additional funding round: 52% moreEach additional funding round
VC backing: 21% lessVC backing
Each additional year to last funding: 16% moreEach additional year to last funding
Each additional investor per round, on average: 15% moreEach additional investor per round, on average
Each additional strong industry relationship: 4% moreEach additional strong industry relationship

Each effect holds the other factors constant. These show associations, not causes.

Technical Foundation

Excel
Full statistical battery: odds ratios, chi-square, t-tests, ANOVA, MANOVA, regression — 28 reproducible sheets.
Data cleaning
Range audit, external verification of contradictory dates, explicit placeholder coding for missing values.
Regression
Built and validated the 7-predictor model explaining 65% of the variation in total funding, checked for multicollinearity (VIF < 2) and cross-validated (R² = 0.63).
Exploratory Data Analysis (EDA)
Checked skew, outliers, and distribution shape on every variable before running any test.
Hypothesis Testing
Matched each business question to the right test — odds ratios, chi-square, t-tests, ANOVA, MANOVA — and validated assumptions before trusting the results.