Episode 11: Infuse Analytics Everywhere

Hey listeners, welcome to SphereCast, a podcast all about technology, technology advice, technology inspiration, and how real entrepreneurs have used technology to build their businesses from the ground up. If you're wondering how technology can support your business goals, rest assured, our guests have been there and done that.

Numbers don't lie. These days, companies that have access to the best numbers, the richest data, have an immense advantage over the competition. Thankfully, in today's digital age, those numbers are readily available. All you need are the resources to tap into them and the smarts to make decisions off them.

Hey, it's Luke from the SphereCast, and this week our team interviews Aviad Harell, founder, Chief Operating Officer, and General Manager at Sisense. Sisense is a platform offering users custom analytics that can be embedded into essentially any platform via integration capabilities and open APIs. As the company's mission statement says, Sisense allows users to infuse analytics anywhere, tapping into deep insights to inform better decision making. To date, Sisense has partnered with over 2,000 top companies to make this vision a reality.

As I said earlier, numbers don't lie, and Aviad understands the competitive advantage that quality analytics can bring to an organization. On the podcast, Aviad discusses the birth of his company, how his startup learned from specific mistakes in its earliest days, and how he's learned to differentiate between temporary market trends and longstanding market changes, and why his company decided that openness and hyper customization is best for everyone, which was ultimately the biggest risk they ever took. So without further ado, here's our conversation with Aviad Harell.

Hi, my name is David from Sphere Partners. I'm really happy to have Aviad Harell here, co-founder of Sisense. And with us is also Mario Schwartz from Sphere Partners. Hi to both of you. I'm very happy to have you here on this very special episode of SphereCast.

Hi David, thank you for inviting us.

Hello David.

Aviad, we'd like to hear a little bit about your journey as a co-founder of a successful startup. Can you please tell our listeners what inspired you to join and co-found a company?

Hi everyone. As David mentioned, my name is Aviad Harell, I'm the co-founder, one of the co-founders of Sisense. The story of the beginning of Sisense is almost a cliché. We met, five of us, in college, in computer science school in IDC in Herzliya in Israel. We partnered and became very good friends. On our last year on college, we joined together into a course called open course in which—it's a yearlong course in which you are supposed to group together with a group of students and choose an industry, choose a problem in that industry, and suggest a solution for that problem.

Me and my co-founders and my friends grouped together, and we chose the BI, the analytics industry. The problem we wanted to tackle is the cumbersome of the existing solutions back then that ruled the BI market, and this is what we wanted to tackle. During the last year of college, we actually built the first client tool of Sisense. When the year ended and we graduated, we understood that we have something that is bigger than a college course or assignment, and we decided to start Sisense. So I know it's a little bit too typical or too cliché, but this is how we started Sisense.

That's great. Sounds like a great way to meet your future partners. When was that? What time was your kind of first—what year was your first meeting or product? When was it established?

2006, I think. Yeah. A long time ago.

A long time ago, and you've done quite a journey since then. Can you tell us a little bit more about the founding team? So you mentioned those are fellow students from your from your college degree, but what made the different founding teams what they are? What did they contribute?

So the founders of Sisense are myself; Guy Boyangu, which is the current CTO of the company; Eldad Farkash, which is no longer with the company, left about two years ago after a very long run at Sisense, is a very successful entrepreneur, a serial entrepreneur; and Adi Azaria, that also left the company about three years ago, and also a very successful entrepreneur.

I think each of us brought different superpowers to the equation. We have Guy Boyangu that was the super architect of the company, software architect. We had Eldad that was a phenomenal product/technologist evangelist, and we had Adi that was on the operational side of the go-to-market. The combination of all of us together, I think, was the driving force of starting the company and scaling the company part of that journey.

That's great. Sounds like a very well-balanced and very unique set of skills across your founding team. Can you elaborate a little bit more about what is Sisense really, and why did you create that service? I mean, you did mention it initially that you identified a problem within a market, but what made you come up with the way that you designed your solution, and how was it different than other solutions in the market then?

Yeah, great question. So we identified early in the process that back then, and to a vast extent today, the existing solutions that rule the BI industry are pretty good solution, but they all suffer from pain point, a common denominator pain point that they all suffer from. And this is the total cost of ownership.

If you look at the vendors back then, like MicroStrategy and Analysis Services and BusinessObjects and a few others, they were, and still today, great products. But it takes an average of 11 months to implement them in an organization. At the end of that journey, you're prompt with a pretty solid solution, but there is a reason why it took so long to implement them. The collateral impact of that is that if you have something that you haven't thought of initially when you implement that solution, or you want to extend that solution, you can do that, but it's going to cost you a lot of tears, blood, and sweat and money in order to achieve that.

Our approach was to leverage advanced in technology that will allow us to focus on the end user, on the consumer of the BI platform, and to provide them some kind of productivity tool or empowerment tool that will allow them to control the majority part of the BI project. How did we do that? We understood that at that given situation, the end users are only playing with the front end, with the last mile of the analytics. They're playing with charts, they're changing the style of the charts, they may be able to filter it a little bit, but that's it. Whenever they need to address the other part of the iceberg, the underlying data, they do not have mandate to do that, and they address their IT, or they call their IT to help them with that.

We decided to build a platform that will address both sides of the equation, right, the data preparation part and the analytics part. The notion was that if we will provide these two capabilities to the end users in a simple enough way, they will be able to own, truly own, the majority part of the BI solution.

Now, it might be a little bit tricky, right, because there are very sophisticated ETL tools, for example, or very sophisticated data warehousing platforms and tools to help IT department to achieve their goals. This is absolutely not one of the goals of Sisense. When I refer to data preparation capabilities part of the BI solution, I refer to the last mile of the data preparation. If the use case requires heavy-lifting ETL process, you should do that with Information Integration Services by Microsoft, or Talend, or Informatica, or what have you. But for the last mile of data preparation, to create aggregation, to concatenate first name and last name, to create data manipulation that are truly the last mile of the ETL process of the data preparation process, we can empower the end users to do so. And when you do so, you empower them to take bigger parts, bigger chunks of the BI solution.

This is what we had in front of us when we started the company, and this is exactly the platform we built eventually.

Thanks, that's that's very helpful and kind of very well explaining why the need was there and what makes Sisense different from the get-go.

Aviad, the question I would have for you is the typical one that probably you're getting from your clients in the last couple of years, which is how are you adjusting your product to be able and capable of working now with, you know, advanced analytics and machine learning? What is that next step that your product is allowing them, you know? So are you allowing your product to work with, you know, Python, with R, with the new tools that are going to take you to the machine learning world?

This is an excellent, excellent question. What I've described earlier were and still is the main reasoning for Sisense, why we started Sisense. And as I mentioned earlier, we founded Sisense more than a decade ago. And I think one of the biggest challenges is to continue adjust your offering to the evolution of the market, hence your question.

Two years ago, Sisense bought a company called Periscope Data out of San Francisco. Periscope Data is a company that build notebook-like products that allows users to type SQL, or R, or Python and render them into charts in a very easy and straightforward way. The idea behind the acquisition was to enhance Sisense offering with these capabilities. And since then, we've been working very hard to integrate these capabilities throughout the entire value chain of our product. So today, we also allow our customers or our users to enrich their data models with Python and R. We allow them product flows that are more natural to the regular product flows they expect as data engineers as well. And we also introduce them with a lot of integration with platforms like SageMaker and such.

All of that, by the way, we're doing in parallel of enhancing every capability, or as many capabilities with our within our platform with AI smartness, or fairy dust of AI, in a suggestive manner. We want to enhance the experience of our users by introducing AI suggestive engine that will make their work with our platform much, much easier and more guided by the tribal knowledge of how people work with data, how people in their organization work with their data, and so forth and so forth. This brain, we call the knowledge graph, which is a metadata that is being managed internally in our platform, that learns constantly how analytics and operational dashboards are being used with each one of the use cases and each one of our customers, and by that, allowing our users to have a more enhanced experience when they're using the platform.

Perfect. Now, and you just basically setting yourself for the future, which I was asking you the question because of how you said that what drives Sisense is bringing the value on on what they're investing on the BI, and that this value is going to keep on going with your product. It doesn't just stall once you get to a certain point. So that that's a perfect view of where Sisense is looking at the future for their clients. So thank you for that.

Absolutely. And just to add to that a last comment, the current focus of the company is to empower users to infuse analytics everywhere. I know it's a nice slogan, but there is a lot of depth behind that. And the idea is that our platform allows our users to enhance their business flows wherever they are with analytics, with machine learning, with insights. And this is the vision of the company. This is the focus of the company. And in order to to achieve that, we must make sure that we enable our users to use the right tool for the right task, right? If they need Python or R to enrich the data or to analyze the data, in one way, they can do that. If they need a user-friendly drag-and-drop engine for enhancing analytics with business users, they have it. If they want to use their already existing systems and data warehouses, they can do that. If they don't have it, they can use the Sisense data engine in order to achieve their analytics goal. So all in all, the platform allows them to achieve their analytical goals and infuse them wherever these insights are required.

Thank you, Aviad, for that. That basically shows the vision and why people should be looking at products like yours to say, okay, I'm really going with somebody who understands how the data is evolving and how we're going to need it in the future.

Yep. There is another trend, and David, tell me if I'm drifting a little bit—

No, go for it.

Another trend that relates that relates directly to what Mario mentioned, and this is the trend of going beyond dashboards, right? Dashboards are integral part of analytics and BI, right? You want to analyze something, the go-to vision you have in front of your eyes is dashboards. But we believe that the world of analytics is changing. I don't need to leave my existing workflow in order to analyze something and to jump to a dashboard. We want to bring the insights to your existing workflow. So for example, you work with Salesforce, right? You're an account executive and you work with Salesforce. And this is your day-to-day, right? You're within Salesforce pages, and you want to enhance the experience of the account executive by by giving him analytics of the quality of the leads, or the conversion rates of his funnel, and things like that, you can infuse Sisense insights into these workflows versus asking that account executive to leave Salesforce, jump to a dashboard in a different system, analyze something, and go back to Salesforce. So this is the essence of infuse analytics everywhere.

Would you say that this is a different a different name for white labeling your service and integrating it into an existing platform in a seamless manner that doesn't interrupt the user, and basically creates more value on top of an existing service?

It's more about the integration and embeddability of the platform. I mean, white labeling, of course, it's part of it, but this is only the icing on the cake. The ability to embed the entire workflow of analytics into an existing workflow, this is the essence of what I'm talking about. It requires capabilities of white labeling, it requires a lot of integration with many, many systems, it requires componentization of every part of the value change of analytics so you will be able later on to embed it into whatever business flow that you have in your organization.

So I'm going to piggyback on that one, and the question would be how are you deciding to which platforms you're doing this integration? What drives your decisions of I need to, you know, integrate with these platforms? Or are you working with some open source type of code like they do with the Python libraries? What is the strategy?

So again, an excellent question. There are few families of integrations of Sisense, right? One is the data source integration. Which data sources we support in order to connect, to query, to pull data from, and such? This is one set of integrations. We support the vast majority of the well-known ones, right? From SQL Server, Oracle, CSV files, Excels, Snowflake, Redshift, BigQuery, all of them. We also have generic JDBC libraries that allows our users to extend it to their own needs, the connectivity part of the data connectivity part of Sisense. So this is one family of integration.

The other family of integrations that we constantly work on is the embeddability of the platform, right? So we provide SDKs for our customers to embed either visualization or query flows, insights from our platform into their systems, either by iframe, very simple iframe and very simple SSO integration, to very complex Sisense.js JavaScript integration libraries that allow them to embed every part of the platform in a hosting platform as well.

And the last part is, or the last family of integration, is what we refer to as a product strategy: API-first architecture. The platform, Sisense platform as you see it, early, very early on in the process we decided on that strategy that we call API-first, that means that every feature in the product, front end, back end, middle tier, we start by designing the API calls, the building blocks of that feature, and then we use them in order to implement the actual features that you see in the system. But more important than that, we expose these API calls to our customers, to our users, so they can take advantage of them when they embed specific flows of the product into their systems.

So if I summarize it, we have data integration capabilities, we have embedding capabilities, and we have the native APIs within our products that allows our customers to customize existing flows with the product, to extend flows in the product that we do not have, and to solutionize, if there is a word like that, or to automate flows within the platform by using these APIs.

So basically, in short, you future-proof your platform by giving the business not only the ability to have all these multiple integrations and capabilities, but you also tell them, "I'm working with you to make your life easier." So you're still driven by your speed to market, like you explained at the beginning. You want your projects to be projects that can be done within weeks and not months for them to get the value.

Exactly. I think it's a great way to describe it. In many ways, Sisense platform is a tool belt of analytic tools or BI value chain tools. You use the tools within that tool belt that you need, right? The right tool for the right task.

So is that because with this statement that you take place, that Sisense was well-known in certain industries more than others as you were growing up, was this because of how they were able to perceive you? A lot of the your early incursions in the market were driven by, you know, engineering, manufacturing people who understood really quick how your approach was more than the typical financial companies. So how do you evolve your your basically your message to that broader audience from from that expectation they had of you when you began that they were looking at your product as a basically, these guys are just expert on this area?

It's a great question again. So the main driver of growth initially, at the beginning of the journey, was driven by customers, right? We did not approach the centralized BI or IT department. On the contrary, we focused on the end users, on the analyst in the HR department, on the analysts on the IT department, or the product team that wanted to enhance a product with some analytical capabilities. So the approach was very product-led growth, right? Less enterprise-led growth that is manifested by a top-down approach of sales cycles. This is how we—I think the rapid growth of Sisense is contributed to that.

In later years, we also invested a lot of efforts in more traditional outbound enterprise sales practices, right, and work very hard with Fortune 500 companies to see how to serve them, to make sense of the offering for them. It wasn't easy in the beginning, but I think in the last couple of years, there is a mind shift with enterprises, right? The number of enterprises that choose a monolithic BI platform or any platform for that for this discussion is slowly, slowly fading. There aren't many that are doing that. The the notion of delegating the responsibilities and empower the independence of each department is something that we feel very well in enterprises as well, and it's a it's a major major force in our success in enterprises as well.

Would it be would it be right to say that when you started Sisense, you had kind of a BI for the people mindset?

We had one of the the coolest slogans that we had a while ago was, "Simplifying complex data for business people." It was one of our most successful and cool slogans back then, which, I mean, when when you draft a slogan, you need to measure every word in that slogan, and we did. And the idea was that if your data if you don't have any data challenges, I mean, you're working with an Excel or a single table, a simple table, there are other tools that will serve your needs. And if you are very technical and you have a tribe of engineers at your disposal, you have solutions for that. But if you're a business users and your data is complex, Sisense is the is a great platform for you.

I I can definitely see that based on your explanations.

I want to go back to your journey as as a startup company, because some of our listeners are entrepreneurs themselves and they're facing many crossroad crossroad decisions about which platforms they should be using in order to accelerate the growth of their own business. So, you know, whether it's a project management tool or reporting or a ticketing tool, or an ERP or accounting, the you know, the list is very, very long. I was just wondering, you know, when you had to take these decisions and, I guess, you're still taking these decisions today, early on, what were the considerations and why why did you decide to choose one system over another?

So I'm a big believer in a few guidelines when you take these kinds of decision. First, suffer the problem personally, right? For example, we had to choose a ticketing system. I decided back then to be on top of all the email chains of customers that send us support emails because we did not have a system. This is just as an example, it's more than a decade ago. But after I got very intimate with the way we interact with our customers on on support emails and what kind of problems do they have, how do they expect the interaction will be with them, I was knowledgeable enough to understand what we are looking for.

Then I decided to take a decision that is not too binding. Meaning, I decided not to choose a solution that will take me a year to implement and then I cannot withdraw from it. I chose a system that, much like the company, was very agile and iterative, so I can examine or measure the impact of that platform as as we progress. Back then, we chose Zendesk because it fits the needs. I was able to start very small with Zendesk, to buy, I don't remember, 100 bucks a month worth of licenses, and to see how it works. And then later on, I will be obligated to take bigger decision.

So the idea is to, in my mind at least, to start becoming intimate with the problem and how things operates in order to understand the requirement, and then start with almost, you know, a very agile approach of, start with something very small that you can iterate and scale it later on.

That's very helpful, thanks. Did you need to use, you know, external integration companies for some of those services, and and why? Or you had everything done in-house, like no no external help?

So the vast majority of things we've integrated in-house. At some points, we when we had to make a large investment in our Salesforce internal development and growth, we worked with external consultant for a specific project. We're not shy of doing that if it saves us efforts, right, if it makes us more efficient. But if I look if I look back and I examine our approaches, most of the times, we just do it by ourselves. We buy products that help us to be more efficient, but the integration and implementation, in most of the cases, not all of them, were doing by ourselves.

We're fortunate enough that we have a large, very technical teams to allow us or to enable us to do so. But in cases we do not have it, we are we're far from being shy to buy some help.

I think that's the advantage of being a technological startup company. You have a lot of technical talent. That makes a lot of sense.

So, you know, with with the world slowly coming out of the pandemic, can you tell us a little bit how COVID-19 has impacted your business, and what kind of steps have you taken in order to make the most out of the current situation?

So I think pandemic had had impacted Sisense like the majority of the tech industry companies out there, and I'll explain. First, the fact that I cannot meet our people, Sisense people, face-to-face, only on on Zoom or video conferences and such. I'm very happy I was very happy to see that the impact on productivity was not major. It had an impact, by the way, a long-term impact, less on the long-timers that have been with the company before the pandemic and we had some runtime together. I believe it had more impact on people that joined during the last 12 months already. I have a few people that we've hired that I never met, right? And people that report or part of my organization that I met only on Zoom calls and such. It has an impact. I believe we were able to accommodate or to mitigate the gaps, and the productivity was not impacted dramatically.

The other impact of the pandemic was on our customers. We have few customers that were impacted dramatically, right? We have customers from the travel industry, we have customers from the accommodation or the hospitality industry, we have customers that one customer that run museums point of sales that were practically shut down for a year. So it had impacted our customers and thus had some impact on Sisense as well. I am truly happy to see that the world, or at least the the the business world, is slowly but surely getting out of that situation, and even customers that were in in bigger problems or had bigger challenges are recovering very well, and we feel it as well. So I'm grateful for that.

All right, Aviad. I would like to ask you, you know, from your experience and and, you know, it's been it's been quite a successful journey for for you with Sisense, what were the, let's say, most impactful mistakes, and then what would be your tips for someone who's just starting off their technology business? What should they be looking out for, what should they be avoiding?

Yep. So I'll try to avoid corny answers as much as possible. But mistakes, we had tons of them, right? My approach, and it's not mine, I think it's the the right approach of dealing with mistakes, is to do them quickly and recover from them as quickly as possible. We had many mistakes along the way, and I'll tell you about the the biggest one in my in my opinion, of course. And I'm sure we're going to have more mistakes in the future. Again, if we'll be smart enough and quick enough to understand that these are mistakes and to fix them, we'll be in a great, great position.

I think the biggest mistake, holistically, when I think about Sisense, was in the early days. When we started the company as young entrepreneurs, we had the notion of the willing to build the perfect BI platform out there. Now, I'll tell you a secret: there isn't a perfect BI platform out there. There isn't a perfect product out there. There are great products, but far from being perfect.

That became a problem when it started to slow us down. We constantly delayed our go-to-market, our willingness to publicly market our product and allow external customers to use the product because it was yet to be perfect, and we just need to build an extra capability, product flow, or something into the product before we push that platform to the market. And it was a big mistake. And we were dancing around the fire for a couple of years before we actually went out to the market. And you know you know what was the compelling event that pushed us over the the chasm?

We ran out of money. This is what happens, right? We ran out of money, and we went we reached out to our investors, and we told them that we want more money, and they told us, "No. You have to take what you have now to the market and see how the market reacts to that." And this is the biggest lesson that I had as an entrepreneur. Because few things happens happened when we put our our platform in the market. First, we understood how little we know the market is or what are the market requirements, right? We had our vision that was a good vision, but very not aligned with the vision of our customers. And the understanding of the approach of building great products. Great products are not being built or designed in an isolated lab. It is an iterative process that you do with your customers and part of your go-to-market approach. And this is something that we learned only after a couple of years of, as I mentioned earlier, dancing around the fire, and only when we were pushed to release our product publicly, only then we learned that. And we lost a couple of years, by the way, which is a great lesson to learn from.

Tying up with that, Aviad, and this is probably part of that lesson that you learned, is how do you handle those market fads, you know, that come up, all all these market trends that they keep on, you know, that people get all, you know, so excited about? And you know it's something that's just basically going to last a very short time, but a lot of of the companies are investing on that. How do you make the wise decision of, "Should I invest on this or not invest on this?" Like, you know, like your approach to analytics that you mentioned before, to me it shows a compromise. So you were smart about, "I'm going to give analytics to the people, but I'm going to do it the right way instead of just, yeah, I'm going to give you a platform so everybody can do analytics."

So I I'm not sure that there is a single answer to that, right? Few thoughts around that. First, I believe that new, exciting technologies do not just pop out, right? You hear about them if you're, you know, deep into the market and you are very intertwined with your industry. You hear about new ideas, new startups, new approaches before they formalize into products even, and of course before they formalize as a market approach, okay? First thing.

Secondly, I believe that if you truly understand the market, you have some sense, not the absolute sense, but some sense of which technologies are going to make an impact and which technologies are just, you know, a thing that is going to fade away, right? Now, I ain't no prophet, but when you truly know your industry, you have a sense of these kinds of things up to up to a constant, as we say.

One good example is Hadoop, right? I don't know if you remember, like five years ago, six years ago, Hadoop was the thing. Everyone talked about Hadoop, right? If you have an analytics solution that does not have a very well integration with Hadoop, you have no reason to to exist in the market and such. And today you hardly hear about Hadoop because Hadoop was an interesting approach for very skilled, very large enterprises to handle, to maintain, and even that was with a lot of difficulties and challenges. And it felt to us back then that it's not maintainable. There are very few companies that can maintain true, working Hadoop integration or implementation in their organization, and the majority of the market will not be able to truly adopt it. Thus, a new way to deal with big data will emerge. And we see we have now the cloud data warehouses, or they even the manage Hadoop in the cloud because people cannot manage it by themselves, or the majority of the companies cannot do it by themselves, and practically made Hadoop integration irrelevant. Now, this is a good indication we got, or a good decision we took back then to not ride the wave of Hadoop.

We also had some misses. The cloud data warehouses wave that is, rightfully so, driving the market of analytics and data, we are very well intertwined with that and we're integrated with all of them. But we could have done that integration earlier if we could understand that the the majority or the size of the impact of these technologies will bring into the market.

That's great. Yeah, actually, I wanted to ask you, I think you pretty much answered it, you know, what was your most successful business risk decision? When you have to determine a strategy, whether we're going that way or the other way, where we're going with this type of integration or the other type of integration, what was the, let's say, risky business decision that you're most proud of that you you you took, you took that you had the right foresight, you had the right understanding and the vision of the market, and you said, you know, that was a really, really successful decision for Sisense, and I'm really happy we took it?

I think the biggest one, the biggest risk that rewarded us tremendously was the openness of our platform, the API approach, right? The fact that we allowed or gave an entry to our customers to customize everything within the platform was a big decision that it's a double double-way sword, right? If you give them access to the internal of the platform, they can screw things up without knowing a lot of what they're doing. But at the same time, for people with knowledge and experience, it can give a lot of power or empowerment to achieve their analytical needs.

And our approach was that if we will spice the openness of our platform with engineers with our side that helps our customers to harvest the value of the openness, it will be a win-win situation. This is what we did. We built a very big, successful, knowledgeable customer success department with many technical people that helped our customers to understand how to rightfully or in a right way to utilize the openness of our platform without, or and reducing the risk of harming their implementation.

That's very insightful, Aviad. Thank you so much for joining us today and sharing from your own entrepreneurial experience, and of course a lot of insights about using data analytics and BI tools today versus how it was 10 years ago. I'm very, very happy you dedicated the time to spend with us and to educate us and our listeners on all of these super interesting topics. And I wish to say thank you say thank you very much, and I'm sure that both you and Sisense are on the right path for further growth and and success, and we're very happy for you.

Thank you very much for that. Remember to infuse analytics everywhere. This is how businesses are successful these days. And thank you for your time.

A special thanks to our podcast guest this week, and once again to our sponsor, Sphere Partners, for bringing this episode to life. If you enjoyed this episode, drop SphereCast a five-star review on iTunes, and share this content with your network. For any relevant links or notes from this episode, check out our podcast website at www.sphereinc.com/spherecast. And always remember, when you think you can't, technology can. See you next time.