As data scientists continue to become an ever-prominent part of the IT team in any business (not just tech businesses), we’re seeing the emergence of more products and services that are built to cater specifically to them. Periscope Data is a startup that has built a platform for data scientists to create fast but highly detailed and customised visualisations — think Tableau, but for engineers — and it has raised $9.5 million in funding.
The Series A round is led by DFJ’s Josh Stein, a new investor in the company, along with participation from previous backers AngelPad, Susa Ventures and Eric Schmidt’s Innovation Endeavors.
Started in 2012 by ex-Googler Harry Glaser (who is now Periscope’s CEO) and ex-Microsoftie Tom O’Neill, Periscope has a number of other notable early fans who invested in its angel round. The list includes Yahoo CEO Marissa Mayer (who was Glaser’s boss at Google), Ellen Pao, Google Ventures and SV Angel. In a business where success is often predicated on the perfect mix of useful and great product, money and the right connections, it’s a sign of how Periscope Data is on track in nailing all three.
Periscope Data (which Glaser tells me never really has any name confusion with the other well-known Periscope) works by taking raw data from various sources — the most popular Glaser says are MySQL, PostgreSQL, Amazon Redshift (which has a partnership with the startup but does not invest), Salesforce and Microsoft SQL Server — and with a few short commands, lets engineers create graphics to help make better sense of the numbers.
The key also is in how fast the service works: 150 times faster than any other visualisation tool on the market, the company claims. The speed is partly due to how Periscope Data was originally architected: the original startup idea, Glaser tells me, was to build a site where engineers could enter any kind of data query based on data in the cloud and get a fast answer (a kind of “Google for data engineers,” as it were). The visualisation tool was an accidental side project to this, but they found was being used much, much more than the rest.
“We got 2,000 users for the first product, which was very fast and flexible and nothing like that that existed,” Glaser said. “We built the original version of Periscope Data for ourselves to look at the data of how the original service was being used. It was humbling when people turned out to be much more interested in that side project than they were in the main one.”
Periscope Data has up to now largely grown by word of mouth and very little marketing, and even so it’s picked up steam in an impressive way. Of the 200 businesses who have signed up so far, customers are “highly engaged” in the words of DFJ’s Stein, with 90% using it at least once per day: “That’s an insane stat for any business app, so we are betting that even more will have similar data science needs.” Periscope Data’s customer base has grown 10-fold in the last year.
Interestingly, while so much technology today feels like it’s out there to “replace” less efficient human labor, Periscope Data sees itself as serving a different purpose.
“So often companies like ours focused on data science, the pitch they give to companies is that there will never be enough data analysis in the world, but if you buy our software you won’t need to hire as many data scientists,” said Glaser. “But the best companies are hiring whole teams of these analysts to compete. That’s something you will continue to have to do, and you should be giving them the best tools to do their jobs.”
It’s also part of a bigger wave of startups coming out that are also trying to do the same thing, some of which are also, interestingly focusing on visualising all that abstract data. CartoDB, for example, focuses on mapping specifically.
Glaser says there are no immediate plans to add new products at the moment, as this is really the first time that the company will market Periscope Data’s visualisation platform. It’s also going to do some more hiring in its engineering team.