It feels likes eye-popping times for those deep into building the future of visual data experiences. With Looker exiting (-> Google for $3B), Tableau exiting (->SalesForce for $16B), and less public, Periscope & ZoomData exiting, the Graphistry team is experiencing good feelings and key reflections. One of them is… the $16B exits are just a prelude to the next $160B in opportunities.
To be fair, we should’t be too celebratory: 2019/2020 will be bittersweet for many of the number of people involved. The dissolution and digestion period by corporate cloud data giants complicates innovation, adoption, and service for many users. And yet, the selloffs are also validation that make technology builders & thinkers even more excited for the next 100X’ing of the human-in-the-loop data experience.
While definitely not our entire thinking, we wanted to share three hard-earned 100X perspectives around what the above companies, our peers, and Graphistry itself have shown:
1. The BI charting space has grown bigger — 100X? — than we thought. Everyone from security & fraud to sales & marketing to supply chain & finance is facing a growing sprawl of internal+external data sources. And yet, most people still do not benefit from good and easy data tools. Tableau made a 100X leap here by introducing drag-and-drop SQL dashboarding. 15 years later, that vision is still out-of-reach: Periscope’s motto of ‘Write SQL, get chart’ is still an exciting proposition to many users. SalesForce admins will soon be breathing a sigh of relief for their dashboarding. I expect new fun tools to continue to emerge!
2. The potential of improving the human-in-the-loop data experience is bigger than the above, bigger than we originally thought… and the solutions do not look like classic BI dashboarding. Many of our users already have tabular bar charts and search (<3 Splunk & ELK!). Despite that, they still struggle with basic visual discovery, surfacing relationships, visually leveraging ML, sharing favorite queries/workflows, visually automating, and all the other things we must do when working with internal+external data. So: Less low-level SQL query authoring, more smart and computer-assisted visual auto-pilots. When logs can span tons of systems and hundreds/thousands of columns, it’s even more important startups like ourselves and our peers focus on 100X enablers. We don’t need faster horses, but trains, planes, & cars!
Video: GPU cloud visual analytics for instant scale & ML over visual UIs
3. 100X solutions: We had started Graphistry with a post-Tableau data+tech mindset, and more has happened since then.
— GPUs for instant fast data: GPUs client+cloud means every interaction — from moving a slider to drilling in — can benefit from massive scale & smart compute in sub-100ms. The open source movement in particularly is finally emerging around GOAI / Nvidia RAPIDS. You can find many of our thoughts & videos for what OSS GPU client/cloud analytics is enabling. GPU serverless is coming too – it feels like we’re in the early days of Hadoop and Spark!
— Virtual graph for fusion: The data relevant to any individual analysis is increasingly spread across columns/tables/databases/APIs. Multiple technologies are emerging for writing queries that can work over different data sources, such as fusing calls over Splunk with SQL and APIs in one query, and ideas like Virtual Graph for providing unified entities and actions over them. We do a lot here, and love how companies like Microsoft are providing the Enterprise Graph API and Facebook promotes GraphQL for this unified data+action future.
— Graph analytics & visual ML: Data maturity roughly follows time series (raw signal metrics linecharts), events (observability via log search over wide columns), to relationships (correlation, entity resolution, patterns & outliers, prediction, …). The rise of graph analytics, explainable ML, and structured data (graph) deep learning bode well for getting analysts out of log hunt-and-pecking and into thinking. Google’s top DeepMind neural nets team is surprisingly vocal here!
— Easy-mode collaboration & automation. ServiceNow and then UIPath made massive pushes, and in security, auto-response orchestration tools like Phantom Cyber and Demisto showed the need in verticals. But these are generally still too weak or cumbersome for when an analyst wants to explore a question, must tackle an incident, or overall manage their teams data experience. Today’s data/automation/ML black boxes do not support reaching in: they feel more like lawnmowers than Minority Report.
On behalf of the Graphistry community, congrats to everyone ending their startup journeys.
And, more so, for those who see this as the beginning for bringing the next 100X shifts to everyone, we look forward to the next 5-10 years with you!