Eureka! with Adam Greco (Amplitude)

“My journey with digital analytics started in the early 2000s, when I was in charge of the Chicago Mercantile Exchange website,” Adam starts. “Every year, the board of directors asked me to do a complete redesign. Eventually, I started wondering: what are we trying to achieve here? Who is our target audience, and what do we actually want them to do? All of that led me to explore ‘web analytics’. Soon, we began tracking the most popular content, user behavior, etc. We also became one of the first customers of Omniture, which later became Adobe Analytics.”

Adam’s Eureka! moment:

“So much of the world is moving digitally. If you want to understand how people are responding to what you’re making, you need to understand digital analytics.”

20 years later, digital analytics is still a major part of your life. What made you stick around?

“What makes digital analytics so addictive for me is that so much of the world is moving digitally nowadays. 70 to 80% of all customer touchpoints are digital, and the pandemic really accelerated that. If you want to understand how people are responding to what you’re making, you need to understand digital analytics.”

“People used to make business decisions on gut instinct. And I do believe there’s still a place for that, and for creativity. Nike didn’t come up with ‘Just do it’ because of data insights, for example. But even they have invested hugely in digital analytics to figure out how to turn interested people into engaged customers.”

When web analytics and product analytics meet

You’ve been involved with web analytics for most of your career, but recently you joined Amplitude, which is well-known for its commitment to product analytics. That’s quite a shift.

“People have made fun of me in the past because I’ve been doing the same thing for over 20 years. Ironically, many of those same people are now yelling at me for doing something else. You just can’t win (laughs). The bottom line, however, is that I find it important to learn, grow, and try new things.”

“And I genuinely believe that product analytics is the future. Younger generations use apps first, and browsers second. In my own life I’m also experiencing a push towards apps, for example when making doctor’s appointments. More and more websites are turning into products as well: they’re almost like apps in a browser. And then there’s the persistent growth in actual apps as well, of course.”

“What’s more, it seems to me that web analytics products have been stuck in a ‘web paradigm’ and are now playing catch-up. Once I started looking into product analytics, I quickly realized that Amplitude was the vendor that was best positioned to dominate the field. Considering that I’m almost at the end of my career, I thought it would be interesting to try to make a difference in where I believe the future is going.”

“Web analytics products have been stuck in a ‘web paradigm’ and are now playing catch-up.”

“Amplitude got a lot of questions from customers who had bad experiences with marketing analytics. So my main goal here is to help the team figure out how we could solve these problems, and whether web and product analytics could actually merge at one point.”

And do you think they will?

“When websites came around, they were mostly used by marketeers to put their paper brochures online. As a result, the website – and web analytics – are mainly the domain of marketing, with the IT department just cruising along and providing technical support.”

“When the first iPhone was released, every company suddenly needed an app. That’s when the product team was born. As apps became more popular, the product team started to gain more influence. That’s why you now have a split between the marketing team who owns the website, and the product team who owns the app, with both using their own analytics tools. Of course, that’s an incredibly dumb idea: there is only one customer. That’s why, eventually, marketing and product will have to work together.”

Using two different tools is a dumb idea: there's only one customer.

"For vendors like Amplitude, it’s a race to be the first to be able to do everything. And we’re not that far off: I’d say that about 85% of the items in a product analytics tool can be applied to marketing analytics. However, there are certain things that are more unique to web analytics that you’d have to add to product analytics to make it a full solution, like figuring out if people arrived on your website via SEO, paid search, an e-mail, or something else. Product analytics tends to care more about what happens once someone is already in the product. For marketing, once you’re in, you’re instantly forgotten."

“Product analytics cares more about what happens once someone is already in the product. For marketing, once you’re in, you’re instantly forgotten.”

“Fact is, it’s much easier for product analytics teams to add the marketing stuff than for web analytics teams to add product analytics features. The latter would have to change their whole paradigm, move to a more event-based model, and stop worrying so much about page views and sessions. It’s a race, and I really like Amplitude’s chances in it.”

Data democratization and data quality

Another thing that comes up a lot in your talks and articles is ‘data democratization’: the idea that everybody in the organization should have access to data, without any gatekeepers. How does this concept fit into product analytics?

"Data democratization is gaining momentum because it speeds things up, and the push towards it comes mainly from companies that want to experiment and move fast. They want people to learn from data really quickly, and a centralized team always slows things down. But there’s a balance to be struck between speed and quality. When you go too fast, you’re often creating new problems, like people creating multiple versions of the same event, etc."

“There’s a balance to be struck between speed and quality. When you go too fast, you’re often creating new problems.”

"Marketing’s waterfall approach, where you go to a central team and wait for a result simply doesn’t work for product teams. They’re much more comfortable with data, and they want to know the answers now so they can try out new things. As a company, you have to take this into account: you need to understand that things might go out of control and that there can be issues with duplicate stuff."

"What you can’t compromise on, however, is data quality. The collected data needs to be correct. This is where you have to weigh the downside of moving slowly with the downside of someone tracking things the wrong way. For example, when you track an event called ‘form views’, you need to make sure that everyone understands a ‘form view’ in the same way. Is it just opening a form, or does the user have to spend a certain amount of time on it. In a data democratization model, this needs to be absolutely clear to everyone. Otherwise, you run the risk of someone reporting a metric to the CEO that is completely wrong."

“That’s why data democratization requires a ‘trust but verify’ model, where you check what people are doing and put a lot of effort in educating the team on taxonomy. But even then, some organizations would rather deal with a couple of mistakes here and there and go fast instead of waiting for weeks for answers that by then won’t matter anymore. It’s a completely different mindset.”

Last question: how do you talk to friends and family about what you do?

“I explain to them that behind every website or mobile app they visit there’s a team of people who want to make it better, to get you to spend more time on it, buy more, or engage with it more. Every swipe and every click that you make gives them a signal. They don’t really care about you as a person, but when we aggregate data from all users, we can figure out where people are having problems or are being successful, and ensure that users have the best possible digital experience.”

“Data tells us where people are having problems or are being successful, and allows us to create the best possible digital experience.”