Before Nathalie Miller decided to walk away from Instacart, the grocery delivery start-up now worth more than $2 billion, she made a spreadsheet to analyze how much money she was leaving on the table.

She had been Instacart’s 20th employee, managing operations during a period of extremely rapid growth, and the sum could have been huge. She quit anyway. Like many strivers in Silicon Valley, she had a bigger plan: to start her own company that could not only make millions, but also fulfill a mission she believed in.

Her idea was to build a recruiting and hiring website that would make corporate America friendlier to women. She called the site Doxa and persuaded an engineer friend to start it with her. They strategized in borrowed office space and recruited people from other tech companies to work on the project at night, promising full-time salaries as soon as they raised venture capital.

Then, six months later, Ms. Miller discovered that she was pregnant with her first child. A baby wasn’t something she and her husband, Isaak Le, had been planning quite yet, but they were excited. Still, a sliver of doubt began to undercut her confidence about starting her own company. “How high can I stack the cards against myself — a pregnant brown woman?” said Ms. Miller, who is of Vietnamese and European descent.

Silicon Valley can seem like the land of overnight success. College dropouts show up and become billionaires a few years later. Mark Zuckerberg was a billionaire by the time he was 23, and Facebook paid $19 billion for WhatsApp when it was just five years old.

But despite their hold on the imagination, these examples aren’t typical. Much more representative are the companies you never hear about because they never get off the ground.

Venture capital firms invest in 1 percent of the companies that pitch them. The overwhelming majority of entrepreneurs who are funded are white men. Just 1 percent are black, 8 percent are female and 12 percent are Asian, according to data from CB Insights, which tracks private companies and investors. Even for start-ups that raise money, the mortality rate is high, with most dying after an average of 20 months and $1.3 million in financing.

In recent months, the fund-raising atmosphere has cooled as venture capitalists react to the poor stock market performance of some public tech companies and question whether the recent fast pace of investment is sustainable. Venture capitalists are making fewer investments at lower valuations.

“There is this delusion that it’s easy to raise money in Silicon Valley,” said Sam Altman, president of Y Combinator, a mentorship and investment program for start-ups. “Raising money is incredibly hard.”

Nonetheless, Silicon Valley keeps seducing people. For those who come, the risk is worth the potential reward — not just riches, but also the more dreamy belief that tech entrepreneurialism is a force for good. Ms. Miller was not immune.

“I want Doxa to exist in the world,” she told me in July, “because it will make the world a better place.”

A White Male Culture

Ms. Miller, who turns 34 on Sunday, grew up in Berkeley, and after graduating from Harvard, moved to Vietnam. There, she started a microfinance nonprofit, expanding it to 92 employees. When she returned to the United States, she began a doctorate program in sociology at the University of California, Berkeley, but was soon drawn into the tech industry.

She joined Instacart in 2013, and saw some of tech’s biggest challenges firsthand. In a year, the company grew to 120 employees from 20, and had 4,000 independent contractors. Hiring well at that kind of speed, she found, is hard to get right.

Ms. Miller also experienced the challenges of being a woman in tech. One day, a new employee told her that he had ranked the hottest women at the company, and she was No. 1. She reported the comment to managers, and the employee was fired the next day.

Though she was satisfied with Instacart’s response, she knew women in many industries faced much worse, and started ruminating about how she could change workplace culture more broadly.

That is when she came up with Doxa. The name comes from a Greek term that sociologists use to describe the things that people take for granted but that don’t necessarily have to be that way.

The idea was for a site that would collect data about the sorts of things applicants really want to know but don’t ask in job interviews: Do you feel as if co-workers value your opinions? Do people actually take their vacation days? How often do you work nights and weekends? It would offer online personality tests, then use psychometric data to match candidates with employers more effectively. It would make money from companies paying for information about candidates and to post job ads, as they do on LinkedIn. Sam Meder, who had run engineering at a start-up acquired by Pinterest, agreed to start it with her.

Ms. Miller also saw Doxa as a way to address diversity issues. The statistics in technology are particularly bad. On average, 30 percent of employees at big tech companies including Google, Facebook and Apple are women, 5 percent are Latino and 4 percent are black. A diversity industry has sprung up to teach companies how to become more inclusive. Engineers attend training sessions on unconscious bias. Companies are hiring corporate diversity chiefs.

Ms. Miller thought Doxa could help by offering data that research shows is important to women — the gender pay gap and the percentage of women on the executive team — as well as factors like whether women felt they were given credit for their ideas. But first, she needed to raise money.

Venture capitalists, who hold the keys to success in Silicon Valley by providing start-up money, are even more likely to be white and male than tech company employees are. Theirs is an insular business. Most investors accept pitches only from entrepreneurs who come through an introduction, and they tend to finance people who have succeeded before, or who remind them of those who did.

According to a 2014 study published by the National Academy of Sciences, investors prefer pitches by men, particularly attractive men, to those by women, even when the content of the pitch is the same. In addition to studying the results of three entrepreneurial pitch competitions, the researchers conducted two experiments in which a representative sample of working adults heard identical pitches in male and female voices. Sixty-eight percent of people preferred to finance the company when it was pitched by a male voice, while 32 percent chose the female.

Venture firms with female partners are three times as likely to invest in a company with a female chief executive, according to the Diana Project at Babson College. Yet just 6 percent of partners at venture capital firms are women.

Because of investors’ lack of diversity, they do not necessarily understand the value of companies that make products for customers who are not like them, said Tristan Walker, founder and chief executive of Walker & Company, which makes health and beauty products for minorities, including the Bevel razor. On his first pitch, the investor told him she didn’t think shaving irritation was a big enough societal issue.

“All she had to do before she said that was get on the phone with 10 black men or women and eight or nine would have said, ‘This is an issue I’ve had to deal with my entire life,’” Mr. Walker said.

Some investors are working to change this. Venture firms like Kapor Capital were formed specifically to finance start-ups with a diversity mission and to prioritize underrepresented founders. Others fill their partner ranks with people who are not white men.

“Deal flow comes in through people’s network, and if you’re a young woman or a person of color, a lot of times you just might not know who that senior person is at that top-tier firm,” said Joanne Yuan, an associate partner at Cowboy Ventures, whose investment team of four has three women who are minorities. “Our portfolio looks more diverse because we know a lot of people who aren’t white males.”