The scarcity of data scientists is a serious drag on the big data hype machine. One alternative is to endorse the "democratization of data," expanding the analytical capabilities across organizations. Recently, Neil Raden of Hired Brains posted in support of this idea, which he terms "pervasive analytics."But Raden also cautioned that without better change management, training, and software, the democratization of data would remain a pipe dream. Not long after Raden's post came out, I had the chance to interview Alation CEO and co-founder Satyen Sangani, who gave me his take on data democratization, and how Alation intends to disrupt the crowded analytics market.
The shortcomings of analytics solutions led to the founding of Alation
Sangani has seen his share of large scale analytics environments, having spent the better part of a decade at Oracle, where he was responsible for data warehousing, performance management business, in financial services. Selling analytics solutions to the largest financial institutions in the world, Sangani saw first hand how users struggled with these systems once they were installed:
We watched users struggle with trying to find the data to populate our analytical systems; these customers would have thousands of different source systems, which all fed into the applications we would run. This meant lots of manual effort with a team of consultants in order to stand up one of these warehouses. I realized that if it took you three years to get to a point where you could answer a question, that going forward, that wasn't going to cut it.
That struggle-to-value became part of Sangani's inspiration to leave Oracle and co-found Alation three years ago, along with fellow defectors from other industry giants, including PhDs from Google and a designer from Apple. Alation officially launched on March 31, 2015, but with major customers such as eBay, Square, and big financial institutions already on board, it's fair to say that Alation's enterprise push is on.
Alation's self-described goal is to "centralize knowledge about your company's data and how to use it, unlocking the unrealized insights in your data systems." But why now, and not five or ten years ago? Part of Sangani's answer is the obvious: the proliferation of data (and data sources). But Sangani also pointed to new ways of monetizing insights and predicting behaviors:
Today, it's about monetizing all that data through the generation of insights. Whether it's LinkedIn's data products, or banking products where they can give you the best credit opportunities, or the best credit card to apply for. Or: think retail for loyalty program offers. In every one of those cases, there is an underlying data component.
But to get there, the enterprise must know more about the customer - not just the transactional information of the past, but actual behavioral information. Mining that data leads to revenues. Mining that data leads to increased profitability. Analytics used to be an elective exercise, and largely a backwards looking exercise. Now the businesses that are able to effectively reduce time to insight generate more insights, are able to close deals faster, and generate more revenue. We're part and parcel in that curiosity I think.
Democratize the data, or hunt for data scientists?
There's another problem Alation wants to solve: investing in analytics that too few use:
People are finding that, though they make those analytics investments, a large number of people in the enterprise just don't use the data. Then there's a question of well, why? Why are companies not more data driven? What are those barriers to entry? We are able to address those issues head on.
That raises the democratization of data issue, so I asked Sangani for his take on that:
The expectation that somebody needs to go get a PhD to understand to extract data value is certainly one way of addressing the problem. But I think that the world of consumer software and the Internet as a whole has changed that. When you go to LinkedIn, you essentially get as much expertise as any recruiter would have about the individual you are trying to recruit. You can do that literally in a matter of seconds.
That doesn't mean that you've gone and gathered all the information. It means that when you search for that particular person or thing that you're looking for, LinkedIn guided you to the information that you needed, so that you could contextualize what you could learn about.
Everybody doesn't have to be a data scientist. Everybody just needs information in context. The problem is not giving everybody all of the knowledge they might need to understand every aspect of analytics. You just need to give them enough context so that they can determine whether or not they're on the right path or the wrong path - or at least who to ask if they don't know.
In theory, a smarter system would lead to better answers. Sangani believes that Alation literally becomes smarter by cataloguing the know-how of the company's users:
The problem with analytics is that sometimes slightly off is majorly off. You could be like, "Oh I want West Coast profits." Well, West Coast profits could mean profits from California, Oregon, and Washington. But depending on your system, west coast profits could also mean profits from California, Oregon, Washington, Nevada, Hawaii, Alaska.
The reality is that you need that context to understand whether or not you're getting the right information. By building a very sophisticated, very deep catalog when somebody asks a question, we can give them that exact context. We can sort of say, "Hey this is the definition we're working on. This is how it's being used, and this is who else in your company is using it." That has a lot of power.
For Alation, that means finding companies that have already invested in analytics tools, whether its big data Hadoop flavors or visualization tools, and helping them to fuel adoption of those systems with a company-wide knowledge base - extracting know-how both from systems and from the in-memory brains of users.
Putting "time to insight" to the ROI test
Given that Sangani is so vocal about time to insight, I pushed him to define ROI results. How do you measure time to insight in the first place? Sangani:
The problem of knowledge gathering is the underlying problem with analytics. People spend 70 to 80 percent of their time buried in some report, effectively looking for the data, trying to understand the data - Before they can answer a single question.
So if Alation lives up to its promise, it must ultimately reduce that squandered time. Sangani pointed me to their web site, with results that vary based on the customer's use case. Examples: Alation customer Inflection reported a 20 percent increase in productivity by their analysts. Another: MarketShare reported a productivity increase of 50 percent. In closing, Sanganj took the results question one step further, by returning the focus to his own company:
That's why we measure productivity. That's why we ask about number of insights generated and time to insight. While we sell software, fundamentally we're selling a way that people should work. Obviously, we are big believers in the power of data, and the power of what data can do to build a more transparent organization.
It's about building organizations that can move faster, that can be more responsive, and so we measure those things. We measure those things in our customer interview processes, and in our implementations, because we want to make sure that's what we're tracking too. It's not whether or not we've installed some software, it's about whether it makes a difference in peoples lives.
Image credit: Group of workers people. © Kurhan - Fotolia
Disclosure: diginomica has no financial relationship with Alation. Alation's PR group reached out to me, and I found the topic of interest.