How Self-Service Analytics Creates a Major Shift In Product Mindset

I've recently come across a statement in the media that the next level of business value can only be generated by bringing in non-data experts. This has caused me to reflect on data democratization and its role in unlocking this new value.

I sat together (virtually) with , CEO at , to better understand the evolution and consumerization of data analytics and what the future holds for it.

Before joining Mixpanel, Amir served as CFO at Spiceworks Inc., an online where users can collaborate and seek advice from one another, and also engage in a marketplace to purchase IT-related services. Previously, he held various leadership roles at Twitter Inc., where, according to , he helped scale the company from 150 to 4,000+ employees and also helped lead the company through its IPO.

In our talk, we discussed why self-service analytics is here to stay as a driver of product-led growth (PLG), how new-gen SaaS companies make a difference and disrupt the user experience, how new market entrants can survive and thrive in a highly competitive environment, whether Twitter is still an amazing product, and other thought-provoking stuff.

Enjoy reading!

Amir, how did we achieve the level of data democratization we have now? What was the evolution?

For a very long time, this world was eaten up by spreadsheets. With Microsoft Excel being dominant, you'd still have very few people in an organization that really understood and could model complex spreadsheets.

And as things progressed, we went from spreadsheets to different data storages, tools, and sets of databases. We needed BI tools and SQL to join those datasets across various databases and then roll up dashboards to share with the company leadership. 

So, when it comes to data analysis, it remained a prerogative of data scientists and engineers for a while before a very constrained data model for marketing analytics was built and served for some time by Google Analytics and other platforms.

And then, a very thin layer of functional analytics based on SaaS tools was added, which has taken up the economy over the last 10 years. 

What we now have left is the worst possible scenario, where we have a set of very complex tools in BI, tools that most people can't access if they want to understand data in its entirety. 

And then we got to the point where we had a deep level of a silo, with Salesforce serving the sales teams, Marketo serving the marketing teams – multiple functional analytics tools serving an incredibly thin level of insight into their respective functions and failing to provide the full context of the customer journey.

These things are self-fulfilling prophecies when you force people into these disparate tools that don't communicate with each other. The insights people draw and how they can see their impact on business growth are limited to that narrow view.

And then you have this incredible dynamic where you go to a quarterly business review, and the marketing team is celebrating; they're super happy because their local metrics are doing amazing. The product team is celebrating. Everybody is happy, but the revenue is down. 

And now, thanks to emerging tech, we're entering the age of event-based self-service analytics able to provide an incredible depth of insights that people with no data engineering background can easily access, leverage, and convert into value.

This is an entirely new world where we get to see our impact on things that matter. Things that matter are ultimately tied to the company's outcomes in driving growth and monetization.

So, is it fair to say that cloud and artificial intelligence advancements have become the main drivers of data democratization?

The move to the cloud had essentially two stages.

Stage one was SaaS that just copies everything on-prem and puts it on the cloud, which is Salesforce, for instance, or Workday. SaaS applications are mirror copies of their on-prem equivalents served on a cloud. 

We're in Stage two now, with product-led growth companies like Mixpanel, Slack, Notion or Figma taking center stage.

At their core, their products innovate on user experience, workflow productivity, collaboration, and data connectivity. These tools are now defining and ushering in the new era of data democratization and disrupting "traditional" user experiences. 

At Mixpanel, we provide powerful self-service product analytics, allowing other companies to build better products by having a holistic understanding of how exactly people use and interact with their products. We make data accessible to anyone by generating interactive reports, letting you query your data with just a few clicks and then see visualizations in seconds. 

The magic of our product is really in the fact that we have a very simple, straightforward user interface that is built on an event-based model, which is the purest and most raw form of data. 

Interestingly, you and l process the world in the same way. We process things that happen around us or a set of actions that occur over time.

And that's exactly what event analytics lets you shape, analyze, and segment. And it's incredibly generic! Everything can be modeled in this way.

The simplicity and the power that event-based analytics creates are the approachability to the UX, and the ability to draw insights and build incredibly complex metrics that are very hard to compose in SQL and graph in Looker and Tableau.

You can instantly generate those insights in Mixpanel with a little user learning curve. You learn a few components of what an event is, what an event property is, and who is the user that's taking actions. Then you can create cohorts, you can look at user behavior over time, you can look at retention behavior, conversion behavior – you can draw deeply intuitive insights in the same way that users think about processing the real world. 

That is, at its core, the future of analytics. 

While self-service analytics helps mitigate the silo approach, can it also help promote best practices around trust and credibility? 

I love that you bring that up! You can certainly self-service something, but if it's not trustworthy or credible, then there's no point and no use to it. I believe the credibility story on why data is becoming more and more reliable and why we're able to give people the ability to get to that level of accuracy has largely to do with this amazing ecosystem that has been built over the last ten years around cloud data warehousing.

All the governance tools that have come to life have transformed how you can manage your data, helping increase trust and credibility.

These are things that most of us wish we had a long, long time ago. After a decade of learning, we're here with all the different companies that have enabled this. 

Amir, how do you define product-led growth?

In the enterprise world, you had this fundamental truth for a very long time that it was incredibly difficult to procure and implement on-prem software. You would sign 5-10 year contracts because of just the sheer complexity of getting something like that rolled out.

As I've mentioned above, we had a gradual transition to the cloud. At first, SaaS companies took off, making it super easy for anyone to get up and running instantaneously. 

And while it took time for enterprise companies to slowly gravitate and realize there's a better way of doing business, a new generation of companies with innovative products in different categories emerged, allowing for a completely different dynamic of switching costs and deployment.

The hurdle went down, and then right next to that, we had another cultural change from a very top-down to a very flat decision-making. 

In the past, all decisions were made by the C Suite. And then, there was a moment when the knowledge economy became the dominant force. As talents became the most valuable assets for any organization, business leaders asked themselves, "How do we recruit and retain the best people in the industry?" "How can we give them better tools so they can get their job done even better?"

Did you know that PwC had used Lotus Notes for a very long time because IBM was a big client of theirs? So, they used the IBM product until around 2010-2011, when they skipped Microsoft Office and went straight to Google Docs. It was like skipping desktop to go to mobile – such a crazy shift. All to satisfy employees and give them more flexible and robust tools.

The product has to rule. The product has to rock. Otherwise, you have no chance.

Suppose you try to enter a new, incredibly competitive market. In that case, the only way you're going to win is by having a great product. 

So, in the enterprise, we've mentally taken ourselves away from this super legacy way of thinking and have come up with product-led growth. It's just like the consumer business.

At the end of the day, users will decide whether your product is worth buying or keeping. And in the new SaaS world, there is no other way to live and survive than to have an amazing product.

You can call it product-led growth; you can call it consumerization of the enterprise; you can call it whatever you want. You should have a phenomenal product with really happy end users. Period.

And what makes you confident Mixpanel is an amazing product? What's your key metric?

How do we know we're building an amazing product? 

At Mixpanel, we've reached the point when we told ourselves we're going to use Net Promoter Score as our key metric for the company. We're going to look at what our end users are telling us about how we're doing, and if we are not getting the love, then we have problems. We had to keep going and grinding, right?

And we did that and went from single digits to double digits to where we're now with a consumer-grade Net Promoter Score for how our end users enjoy using the product. And that became the catalyst for product change based on what our users wanted.

While analyzing user feedback, we came up with many discoveries. One of them is that no one wants gated content; no one wants to fill out a form to read a document. The same refers to pricing. After we realized users loved transparent pricing, we set ourselves off on a two-year journey to make our pricing more and more .

Have you ever dreamed of going into an Apple Store and negotiating the iPhone price with the salesperson? Why should you place any custom requests and go through several rounds of calls and meetings to determine the final product price?

Long story short, nobody wants to negotiate. Negotiation is a means to a very bitter end. And that means that in the end, even if you got the best price, you still don't know if you got the best price.

You walk away feeling like, man, maybe we could have gotten more. You walk away feeling hurt because you realize you are subsidizing the good or bad deal that other people have gotten, which is insane.

So we came out and said – no more. We're going to give the best price that has the best margin integrity to everybody. We're not trying to haggle or charge this customer more than that customer. We're just going to make it an incredible buying journey, and we're going to get rid of this friction in the buying process.

If you're just trying to buy the tool and get up and running, the last thing you want is this information asymmetry. You don't even know if you're going to be able to use this thing. But companies won't give you the price. Like, oh, no, we can't talk about price until our fourth meeting. And what is the most valuable thing in your life? It's your time. Give me four more hours, and I'll tell you how much it costs.

We beg to differ.

So after you implemented this transparent pricing system, how did it affect your sign-ups and bottom line in general? 

It's still early days for conclusions because we're barely out of the gate with this. Still, the hypothesis we had while making this change regarding just how much of a better conversion this would have to monetization is well validated so far in our limited data.

Thanks, Amir. Now let's talk about how customers use self-service analytics tools like Mixpanel? What're the typical use cases?

A few years ago, this would have been a very focused conversation around product development like, oh, Mixpanel customer XYZ did that and then changed the UX and improved the version by some crazy amount, right? That would have been the case study that I would have shared with you. 

But the truth is today, people are using it for all sorts of stuff. We have a recruiting funnel running in Mixpanel. BI companies use our self-service analytics in their PLG sales business. One of the largest coffee retail companies uses us for their marketing lifecycle, to get promoters and improve loyalty adoption. In addition, we have these massive tech companies where the founders use Mixpanel every day to understand their application adoption and engagement metrics.

I was talking to a CFO of one of our customers the other day who said he used the leading indicators in the product data to better forecast finances and to provide those leading indicators to his investors for decision-making. That being said, self-service analytics use cases are limitless.

It's powerful and it's just beginning to find that foundation. And what makes Mixpanel powerful is the order of magnitude.

With cloud data warehousing taking off, adoption becomes more massive.

By the way, what's the size of your social engineering team? Is it distributed?

About 100 specialists today. We are fairly distributed, but mostly in the United States right now. 

And what are your plans for the future? Where would you like to see Mixpanel two to four years from now? 

My mission has always been helping other businesses scale and succeed no matter where I've worked.

One thing is true – as soon as you get to a certain level of complexity, most people just don't know what's going on. 

Assume the company has a successful product, and it takes off. And then you make a bunch of new bets. You roll out new products, you launch new features, and then you go out and ask if they're making a difference. Are we just enjoying the same wave that took us off in the first place? Is the new stuff making a difference to us or not? 

And you would be shocked at just how much gap there is between awareness of what is actually making a difference and making an actual impact. 

Event-based analytics, in the way that we are pursuing what we're innovating on, is core power. It's that very attribution that can tie in the core events and outcomes shaping your company and tie that to what made the difference. Was it your marketing campaign? Was it the product shift? Was it some new positioning you launched? Was it a new geography that you tried to penetrate? 

These factors and nuances typically go into a company's monetization and growth. And thus far, all of the limitations that we talked about at the beginning of this call are really preventing companies from their ability to make smart, informed decisions about which investments to continue, which investments to stop or what new things they should go after based on an understanding of their actual revenue funnel.

And that is where we're going to be. Hopefully, our vision will be realized, and Mixpanel will be able to genuinely impact how people's work directly ties into an outcome that drives the company forward.

Amir, I'm also curious to learn more about your time at Twitter. So, you'd spent several years at Twitter in different roles. Could you imagine back then Twitter would eventually become what it is now?

Twitter is really complicated, and part of the complication had a lot to do with its public nature. It was a closed loop because Twitter was such a live, public-facing social platform, generating a lot more opinions and public interest than even Facebook which was significantly larger and more widely adopted. In many ways, Twitter's power was indisputable in terms of what it did, how it shaped the world and how it progressed to have the impact that it still has nowadays. 

In the early days, we had two core paths that we were going to take to build Twitter. One was a protocol – the pipeline that fuels all the different applications that leverage this one-to-many technology.

And the second way was for us to go out and own the end-user experience. And the moment we made a decision to own the end user experience had real consequences in a way that now we had to govern that experience. We had to have all of the different policies and features that have become more and more publicly relevant as the political climate has changed.

From that moment on, Twitter was no longer just a technology company.

It's now at the heart of how information is being exchanged, how trends are being set, and how people shape their opinions around different topics. In the early days, scaling something like this was an infrastructure problem. 

Now it's as much of a governance and a human problem as a technology problem. My take is that I don't particularly have any opinion that I'd like to share about the cycle of changes that Twitter has gone through from an ownership standpoint over the past years.

But as long as we stay true to the real impact that not just Twitter, but all social media companies have on people and society at large, these products can be incredibly useful and valuable if we take those things seriously.

Any powerful technology should also be thoughtful about how it should scale and shape to help build a better society. How we get there is more than one lane. Facebook has taken a very different path to get there than Twitter.

And it's okay that there are different ways because then we'll see if one works and we can replicate it on the other. I'm glad they're not just copying each other on this route.

Did you like the interview? Please share it to help more people read it.