Ackerman: Data Activism

15 minutes, 6 links
From

editione1.0.2

Updated February 11, 2023

Maya Ackerman (WaveAI, Santa Clara University)

Maya Ackerman is at first sight an unlikely activist in the space of DEI in startups; AI professor at Santa Clara and founder of a music-generating AI startup herself, she turned to investigate how bad the lack of diversity really was some years ago based on her own experience as a founder trying to fundraise. Very quickly, she started poking well-researched holes in the bad data we keep citing, and has maintained a steady production of powerful weapons with more precise data and insights ever since.

Interviewed June 2021

Machine Learning for Investors

Johannes Lenhard (JL): You are a computer engineer and an expert in artificial intelligence and computational creativity. How did you start to think about AI and startups in VC?

Maya Ackerman (MA): I started a startup through some of my work in computational creativity, particularly in helping people to write songs. On the side, I am an opera singer. I did some research on the automatic composition of vocal melodies to help me write songs. After three years, it became clear that this was to be a company so we could share this knowledge with other people. [The company became WaveAI.]

I have had plenty of shocking experiences, as far as biases, that left me extremely perplexed. You never know if it is your gender, and I am an academic so I would never take a sample of one as a serious data set. One of my students wanted to build models to help VCs make decisions. I thought, why not? Let’s pick a little team. We ended up having a team of three students to build machine learning models to help investors make intelligent decisions. We then wanted to look at race and gender, and once we tapped it, it all started oozing out. It is crazy. It is awful. It is shocking.

Data and Bias Exposure

Erika Brodnock (EB): How important would you say the data and the right kind of data still is in this field? Do you think we know everything there is to know already? Is there more to learn?

MA: It is strange to realize how misguided the whole field is, to the point that you think it is on purpose. This level of ignorance can hardly be accidental. Given how many smart people are in venture, there is no way that I am the first one to think about it. They keep looking at totals. For example, they may say, “female-only founding teams dropped from 2.7% in 2019.” This is important information. But, it is easy to attack it with the “pipeline problem.” A critic may argue that not enough women are trying to raise money, and this is why such a small percentage goes to them. We need a better way to expose the bias.

Unlock expert knowledge.
Learn in depth. Get instant, lifetime access to the entire book. Plus online resources and future updates.
Now Available

Instead, let us not talk about totals or the percentage of total funding allocated to women, let’s talk about averages.

On average, if a man and a woman go to raise, how much is each one expected to raise? This is not complicated mathematics. You see the big gaps in the research I did, and I can show you the charts.

···

A woman raises at least 10 times less under most circumstances. How do we go about justifying it? It does not matter how many of them there are, and the pipeline problem becomes irrelevant. Then we can look at education, prior exits, and gender. The data still shows it. The amount of bias we are dealing with here is at a different magnitude than people like to recognize. Another issue with looking at totals is if you look at the total amount of money going to women, at early stage versus late stage, you see an even smaller percent of the total pie is going to women in late stage. Then we say, let us make sure we help there. That is super ignorant, because there are fewer people at the late stage. Instead, if you just look at averages for women, there is about a 35% discrepancy at the late stage and a 65% discrepancy at the early stage.

Current Solutions and Misunderstanding Bias

JL: You looked at the impact of COVID on funding, particularly for female founders. What did you learn from that? What are the nuances that people had not understood before?

MA: The COVID analysis seemed to suggest that things just got worse. There were the same problems that we had before, but worse. It is even harder for women to raise funding than before.

JL: Are the strategies for change drawn from your analysis focused on gender and race?

MA: Current solutions are not designed around what is happening. I have a very controversial view around the birds of a feather analysis, where we say investors like to invest in people like themselves. This narrative needs to either be significantly altered or dropped. It is causing a lot of problems. Firstly, it seems to suggest that we need to solve the VC diversity problem before we solve the entrepreneurial problem, which means that we are looking at decades in the future, and it is shoving the problem aside. Secondly, women are biased against women. There is a study by the UN showing that they are less biased against women than men, but it is still very significant.

People are not born with sexism, but it is something you learn from society. Men and women learn to discriminate against women in certain contexts. This is not criticism of female VCs. Many of them are doing a fantastic job, particularly in efforts to reduce bias, and we definitely need more of them. The issue is about how bias works. We all soak it in from culture. All of us need to work to overcome it. By saying we just need to hire more women, men think it is not their problem. They hire a woman and then she is responsible for the diversity investments, which have a tiny fraction of the money. This is a fundamental misunderstanding on how bias works and how bias needs to be resolved.

But we need to tackle both bias against female investors, and bias against female entrepreneurs. Not first one, then the other. We need both male and female investors to take responsibility for and actively work on reducing bias in venture funds allocation.

By contrast, we look at bias in academia. I am a computer science professor where there is plenty of bias against people like me in this space, but there are male professors who actively work to recruit women. That is part of the reason why we are making some progress in that space. In some spaces, female professors are twice as likely to get a job, because there was so much effort to try to correct the bias.

There are other problems with the current solutions. A lot of venture firms have an explicit mandate to invest in women, and when you look at the details of the mandate, they say to invest in companies that have at least one female founder. Companies that have at least one female founder typically outraise companies that do not have a female founder. The key aspect is who is the CEO. If a guy is a CEO, this company that had the female co-founder was already doing better and they do not need help. They do not need diversity investors. Right. This misunderstanding on the details is causing money that is supposed to help to not help anybody.

The Challenges of Examining Race

EB: Let us look at ethnicity. How is that a bigger issue? What is preventing you from looking at it now yourself? What do you expect to find there?

MA: We have done some preliminary analysis and there are so many pieces. First, this is not about white and non-white people. This is a complex issue. And we need to be careful to not overgeneralize. For example, here in the US, people from different Asian countries get treated very differently when it comes to fundraising. We started looking specifically at Black founders. The problem is that we ran it on data that was not complete enough. The data seemed to suggest that the problem is of a similar magnitude to gender discrimination. Again, it might be even worse, and it might manifest differently. We need big numbers. There is no data. There is a lot of missing data. We do not know if the percentages are correct. We are trying to build our own.

It is so tricky because we need to classify ethnicity based on pictures, for example. We are using other people’s algorithms. It is a very complicated thing. A lot more people can be working on this. They are scared, because you can share some information, and then investors can try to misuse it. The ecosystem is so biased, and we are hundreds of years behind in the venture space. Trying to do something positive is so complicated and you must be so careful. There are researchers who want to work on this who do not because they are scared.

Improving an Industry That’s Decades Behind

EB: Why do you think we are so far behind in the venture space?

MA: There is so much power in there. If we look at how progress is made, for example, in women’s rights, it’s gradual. Venture runs the world to a large extent. There are more gates because it is such an important space. From a sociology perspective, I am not sure how this has manifested. Female doctors or professors, there was resistance to it, but eventually we got there. Now it is the government and businesses, and they do not want us there.

JL: What are some of the big unanswered questions? What is next?

MA: Let us start with the assumption that we do not know anything. I spent a year doing research on this and many things that are taken for granted in venture are just flat out wrong. We need to do analysis for each race carefully. We need to do intersectional analysis very carefully on large data sets. We need to then understand, how does my bias manifest against women in female CEOs versus male CEOs? There is some research and people who are doing good work, but just not enough basic questions are getting answered. If you want to design solutions, you need to understand what is happening. The way it is going right now, we are just relying on the passage of time more than anything else. Certain things may take a very, very long time—and perhaps may never be fixed through a passive or poorly-informed approach. I do not know if there is enough will to really correct it, and if there is enough will to fund research like this.

EB: What are some of the biggest structural issues that you found through your research and the lack of funding for it that prevents real change from happening?

MA: It is an incentive issue. The people in power who are extremely powerful and wealthy, and they do not want things to change. I can easily see for decades, they are going to pay lip service to it and they are going to throw these little funds together, treating them as a charity. Things will go on as they did, most likely. Government is powerless against them; they have laws that prevent discrimination lawsuits, and they are not allowed to do anything that would hurt their bottom line. They can claim that diversity investment can hurt their bottom line, and there is nothing anybody can do about it. In venture, they can have any reason under the sun, and there is just completely no retribution. Somebody in power needs to be willing to take a stand. I do not know if anybody has power against these people. In the States, at least, it is all about money. That is where the money is at. I am sure they have more arms in the government than anybody else, so good luck changing any laws. One hope is to do the research and come up with very pragmatic, narrow ideas to start to untangle the discrimination. Even that is turning out to be trickier than I expected, which is surprising me. There are mechanisms that make this complicated, but it is more promising than hoping for something else.

Corzine: Nasdaq’s Non-profit

Nicola Corzine (Nasdaq Entrepreneurial Center)

Nicola Corzine, the executive director of the Nasdaq Entrepreneurial Center, has been leading the Center’s activities since its inception in 2015. She is leading the Center’s strong focus on equitable entrepreneurship and shares in this conversation what can be done with research, teaching programs, and community to create improvements on both sides of the Atlantic.

You’re reading a preview of an online book. Buy it now for lifetime access to expert knowledge, including future updates.
If you found this post worthwhile, please share!