On the latest episode of the Change Agents podcast, Alter Agents Chief Strategy Officer Devora Rogers is joined by Aarti Bhaskaran, Global Head of Research and Insights at Snapchat. Aarti shares the early-career moment that convinced her research could change minds, what she believes separates a good researcher from a great one, and how she’s thinking about AI’s role in the future of the industry — from easing long-standing bottlenecks to the risks of losing human judgment along the way.

In your research career, what is the insight that changed everything?

I started my career in India, at a time when a lot of brands were entering the market. We were doing a project for Microsoft, who wanted to introduce an affordable, CPU-less computer with a centralized server — so instead of spending a lot of money on a desktop, people would pay a subscription. This was probably my second project ever, fresh out of university, working with the most amazing manager, who taught me everything I know about research.

We did a full study — a market estimation and concept evaluation — and presented the findings to the marketing team and the product lead, who was a skeptic. I think he’d been strong-armed into the research by his marketing team, because he thought he knew everything. We’d done a proper estimation of the market, the universe, the conversion rate — and we told him: based on the research, your conversion rate looks like about one in eight, but in reality you have to cut that by a quarter, so it’ll be closer to one in 32. He thanked us for the findings and basically dismissed us.

We’d put so much time and effort into that report that I looked at my manager and said, “Is this it? Is this research?” She said, “Wait and see.” About six months later, we got another brief from the same company — this time to do churn analysis, because they’d launched the product and gotten subscribers. We did the analysis, presented it, and the same product lead walked in and said, “Before we start, I wanted to tell you — you guys were right. We sold one out of every 30 pitches, almost exactly what you predicted. And when people said no, the reasons were exactly what you said would come up.”

That’s when I realized we are the true voice of the consumer. We help people connect what they’re selling to what someone is buying, what they’re feeling to what they’re thinking, what they’re creating to what they’re consuming. People are complex, and no two people — or two projects — are alike. There are always skeptics in this industry; someone will always say, “this is just survey numbers, I know my product better.” But you don’t know every consumer — you know the consumer through your own eyes. That moment is still one that makes me appreciate the industry we’re in.

It sounds like that might be the moment that turned you into a researcher.

Definitely turned me into a believer.

More recently, there’s another moment that stands out — a project we worked on with Alter Agents, across eight to 10 markets, looking at how people consume content and different platforms, and what makes each platform unique. I went in thinking I knew the markets and what to expect. For me, the best insight is often one that surprises you and makes you question your preconceived notions.

On the surface, everything looked fine. But when we started looking at people who use Snapchat for connection versus how they think about the brand, a relationship emerged. When people don’t say something directly, you can still see it in how they respond to the questions and the content you show them.

What would you say to someone who isn’t initiated in research, or who’s a non-believer? What do you want people to know about research?

Just like any other profession, there’s a lot of stereotyping around research. When I started, people would say, “Oh, you’re the one who knocks on doors and does surveys?” That’s just a small part of what we do. Researchers wear a lot of hats that people don’t fully understand.

Think of the life cycle of even a simple research project: you’re a translator of business outcomes into what you want to learn from the consumer. You’re a study designer who develops complex research designs. You’re a statistician, an analyst, a storyteller, a visualizer, a presenter — and through it all, a translator of the voice of the consumer.

Now it can feel like it’s all boiled down to “just give me a stat.” That’s not what research is. It’s nuanced and rich, and the people who work in this industry are multifaceted — both left- and right-brained. You have to stay aware of every methodology out there and stay informed about the industry through ongoing secondary research. I always tell people: stop thinking of researchers as one-dimensional. We should be the secret weapon in your arsenal.

What do you think makes a good researcher?

It depends a bit on what stage of your career you’re in, but some general qualities hold true. Curiosity is a huge asset — the curiosity to ask why. When we used to get a brief, if the product was in a store, we’d go visit the store to see it. Even now, if I get a satisfaction survey, I’ll take it just to see how the questions are asked, because that’s the kind of person I am.

Storytelling is another one — and it’s especially important now, in the world of AI. Skepticism matters too. You can’t take things at face value, and that’s a skill I think is at risk of eroding with AI, because people take what it produces at face value. You need to ask: does this really make sense? Is it really that positive, or that bad? Does this sit well with what I know about this culture or this country? You need a healthy dose of skepticism.

And the last one is empathy — empathy for the consumer, but also for your client, your stakeholder, or whoever you’re solving the problem for, and for whoever you’re presenting to, which might be a different person entirely. I think empathy is also getting pushed aside in our quest for speed and “now-ness.”

As we look at things like AI-moderated interviews, moderators have historically treated the role almost as a sacred vow — creating a space where people feel safe to open up. Now we’re exploring having people talk to a machine instead. How are you thinking about that shift?

There’s definitely going to be a balance that researchers in the era of AI need to find. Right now, everybody thinks they can “do research” — any input will get you some output, but is it the right outcome? Is it the right interpretation? Are you designing it the right way? Should you even be doing interviews in the first place?

My favorite methodology really depends on the question I’m trying to answer. Sometimes it’s one question. Sometimes it’s complicated — neuroscience, followed by a quant study, followed by a consumer ethnographic deep dive, because I’m trying to get at something that’s very difficult to quantify.

When you’re scoping a project, do you start with the method or the objective?

That’s a great point, and it connects to what makes a good researcher. I always ask: what decisions are you going to make out of this work? A lot of people tend to be reactive — “we’re testing this creative, here’s the sample size, let’s go.” But you have to stop and ask why you’re testing it, and what decision you’ll actually make from the results.

Sometimes it’s as simple as a debate between path A and path B, and when you look closely, the difference is finer than splitting hair — research isn’t going to help you there, you should just run an A/B test. I’ve fallen into the habit of being reactive myself, because we’re problem solvers by nature. But it’s always better to take a step back and ask what decision you’re actually trying to make.

What advice do you have for researchers, or for brand marketers who work with research, on how to think about our work during a time of massive change?

Think of us as the voice in the back of your head when you’re deciding whether to do something — what am I missing, is what I’m assuming correct, what would consumers say if they were in front of me right now? That’s the role researchers play, and it’s also why trusting and respecting our expertise matters.

Sometimes the role gets relegated to “just give me a stat,” but that number didn’t appear out of nowhere — there’s a lot of thought, rigor, and methodology behind it. We also bring years of understanding the consumer from a lens that’s different from a marketer’s or a product team’s. And we tend to be a bit more detached from the decision-making — not so detached that we’re sitting in an ivory tower, but detached enough to know that the answer we deliver might not be the one you want to hear, because it’s tied to consumer truths. That’s what keeps you from wasting hundreds of thousands of dollars on the wrong business decisions.

What are some of your favorite research methods, and why? If you had to go live on an island and could only bring one method with you, what would it be?

I’m going to choose “other, open-ended” and say that for me, the right methodology really depends on the objective and the question. That said, I started as a mixed-methods researcher — moderation, quant, everything end to end. When I moved from India to Singapore, I did a lot of neuroscience work — facial coding, brainwave analysis. I still feel like a quant at heart; numbers are what I do.

But on every brief, especially with my team, we try to push the envelope a bit — can we add some behavioral data and see if it makes sense? We tend to have a “hero” methodology that’s typically quant, because numbers carry more gravitas. But when you want to bring things to life, that’s when you turn to other methodologies to enhance the story. We’re still very much an evidence-backed society, and it’s easy to display evidence when it’s in numbers — but numbers are one-dimensional. Sometimes your best insight comes from a single open-ended question you threw in almost as an afterthought, and suddenly you need to dig deeper.

To wrap up — how are you thinking about AI and the impact it’s having on research?

That’s something that’s definitely keeping a lot of us up at night. It’s inevitable that AI is going to change the way we work, and I’m trying to find the balance of incorporating it meaningfully without losing the human expertise and connection the work requires. At the end of the day, we’re still translating business objectives from a marketing, product, or comms team, we’re still talking to consumers, and we still have to translate those results and present them back to people. The human connection is still there.

Where AI can help is in easing some of the bottlenecks that have existed in this industry for a long time — the speed of translation, transcription and open-end analysis, desk research before you write a brief, or cutting a long, thorough summary down into something digestible without losing its substance. Or building a question bank you can draw from easily, or, in my case, making the hundreds of reports my team produces accessible across the company. Those are areas where you can lean in and use AI.

I know a lot of people are pushing further — using AI for end-to-end research, AI personas, synthetic samples. I don’t think the industry has strong enough voices yet on how to meaningfully incorporate AI, and that’s something I’m actively thinking through, including developing a clear point of view within my own company. Otherwise, everybody is going to think they can do research — but very few will do it well.