Count AI Beta is here!
Phillip van Blerk is the passionate Head of Data and Analytics at Omnipresent, a series B startup that offers compliance, payroll and benefits options to hire people anywhere on Earth. He recently sat with Count Chief Evangelist Mico Yuk to share their journey towards a strong data-led culture at Omnipresent, and how Count helps his team maintain data quality and enable over 130 self-service authors and data engineers to generate and share insights with ease.
Omnipresent started in a familiar place - spreadsheets. Phillip explained, ‘We generate a ton of data and have massive data needs, but we’d mostly been using spreadsheets and source system data dumps. The obvious problem with spreadsheets is that people can edit anything or share data with people they shouldn’t. But more insidious is that they tend to lock insights up in a single file. Nothing gets shared and now nobody can build on top of my amazing insight.’ It worked great for individual analysts but not for a company looking to build a strong data culture.
So Omnipresent developed a data and analytics strategy to create high quality, centralized data products in support of data analysts located in the business so they stay close to the questions they are trying to answer. ‘We needed to upskill people in the company fast so we set up a Community of Practice with a Slack channel and bi-weekly meetings.’ Then they cast about for a data tool that could enable the kind of open, sharing culture Omnipresent tries to cultivate.
‘We had pockets of Tableau in place and were originally planning to move everyone over to that when I found Count. I thought - this is either too good to be true or it’s something I really want. As soon as I got my hands on it I saw the versatility of a spreadsheet without being rigid like Tableau or Power BI. It has what I like to call ‘safe flexibility’. People can do their own analysis, build visualizations and reports, collaborate in real time with stickies, comments, diagrams - but do it using secured data from our data warehouse that they can’t overwrite.’
Phillip sought input from the business right away to see if his initial hopes for Count were accurate.
‘I shared it with the head of the business improvement team and showed him how Count is actually the easiest way to access data and a great place for writing basic SQL to get a table of results and start playing with them. He went wild and ended up making an amazing canvas that explained how to translate your spreadsheet style thinking to the world of Count and brought it to our community of practice to share. It got the whole room hyped up.’
At Omnipresent, Count is used as much more than a simple visualization tool. ‘The other incredible thing about Count is that once you’ve done an analysis, joined tables together, done some filtering and aggregation, you can take all of that code and literally just right click and say ‘copy compiled SQL’ and deploy that code to dbt or wherever you want, because Count uses the correct SQL syntax for your database. Your logic is never locked within Count like it is in other BI tools.’ Count’s unique ability to collaboratively explore data, build SQL pipelines and easily share to other systems enables Omnipresent to bring teams together to solve problems and build models with unprecedented speed.
Phillip sees the ability to easily discover and share insights as key to building a world class data culture. ‘Imagine an excellent marketing analyst who has discovered exactly what channels to target. In an ideal world those are features feeding an ML model, but that excellent analysis will never escape the spreadsheet. Alternatively you build it in Tableau to distribute the insight broadly but we still can’t use it in our ML because the logic is trapped in the BI layer, no data scientist can do anything with it. And it takes enormous effort to reverse engineer it into a data warehouse.’ With Count, that analysis is easily moved into dbt to provide a bedrock for all sorts of ML and operational processes. This key unlocks data modeling at scale for Omnipresent.
‘So how do you not have to appoint 20 data modelers in your organization to be able to move quickly? The answer is Count, where three or four people at the same time are looking at a canvas and rapidly iterating through hypotheses, checking and visualizing the data, refining the answer to the point where they have exactly what they need. Then you have the SQL which has been co-developed by data engineering, data analysts and subject matter experts and is ready for production deployment. It’s brilliant.’
‘This isn’t theoretical - it's happening right now. We are running a reverse ETL that consumes data from Snowflake, and guess where the modeling was done between the engineering and data teams? In Count. The teams could work together, iterate in real time. It’s a beautiful thing.’
Getting people to use new tech is always a challenge - especially when you’re displacing something as ingrained as a spreadsheet. That’s where the community of practice came in. ‘We tried really hard to proactively engage the right people via the community of practice. All the people regularly attending meetings and asking questions in slack got a lot of love from the data team.’
They were also very opportunistic in pulling people into the fold.
‘I would often say, hey you asked this data question. Would you like to be able to answer this question yourself? Okay let’s take half an hour and I’ll teach you how to do it in Count. And then we relied on the network effect to spread the word.’
Phillip also credits the amazing support he and the Omnipresent team have received from Count with enabling their success.
‘It’s unlike anything I’ve ever seen before. Someone comes to me with an issue, I add the directly to the Count Community slack and they get helped right away. The customer support is ridiculously good!’
Phillip’s plan is to move beyond simple self-service visualization and reporting with Count as a key partner.
‘My dream is that anyone at the organization - not just data people - can say, ‘I have a couple of questions’ and jump into Count, write them down in a text box or sticky notes, do a little bit of drag and drop analysis, and ultimately we can deploy that in dbt. That’s not just self-service analytics, it’s self-service data modeling. And that is huge!’
Count is a secure, fast cloud BI tool that you can try today for free. Reach out to have a chat about leveraging us for collaborative, self-service analytics.
