Milking your warehouse for all its worth

Welcome to the first edition of Hot Seat, where we interview some of our favorite GTM leaders who have data and AI top of mind about all the things they’ve been thinking but not saying out loud.

Our first brave volunteer (he didn’t actually volunteer, we texted him and asked him to do this but still) is one of our awesome advisors Winson — we’ll let you have the satisfaction of lurking his LinkedIn on your own, but he’s done a lot of cool things in GTM operations after starting his career as a product manager and developer.

Welcome, Winson. Would love to hear about your background in your own words. How did you get into this rabbithole?

Yep, so let's see my background is pretty diverse. Started off in different large corporations — Bank of America and Liberty Insurance. I was a product manager and developer and afterward, I transitioned into strategic planning just because I was thinking “Hey, why are companies making these decisions? Why are we making these product changes? Why have we decided to sell and build for these new customers and users?”

For the last five years, I've spent a lot of time in GTM and revenue operations to help different organizations scale; I started off at Adobe and then moved to Amplitude, Rippling, and Vitally. Growing those organizations and having a PM background is very interesting because RevOps used to be individual siloes of marketing, sales, etc., and focusing on the customer buying journey has been the mindset that has made me succeed in the role, so I find myself really passionate about fostering that philosophy across different companies and people in those communities.


Very cool. So you can code, you have a PM background, and you see a lot of new tech. What is your hot take on how companies should be interfacing — or not interfacing with the warehouse?

Yeah, I'd say the warehouse is a means to an end, which is getting analysis and insights to drive operational change and strategy in the business. People usually think “Okay, let's put our data into Snowflake, PostgreSQL, or in Microsoft or Oracle” and you would run queries off of it and hire a data team.

There needs to be operators who actually can dive deep into data and know the data tech stack because when you're scaling, there's no one else who knows this stuff. You can potentially get a data engineer, for example, who knows pipelining and ETL so they can get things in a warehouse, but then you’re left with where to go from there. You kind of are just paralyzed with data in a warehouse and no way to use it.

So being able to know SQL really well is a skill set that's really invaluable for an operator these days, especially with companies cutting costs and trying to find people with dynamic skillsets that enable you to essentially do analysis and to run financial models and create analysis that you can just provide to your stakeholders and then build up dashboards and reports to inform where you are from the KPI and OKR perspective.

So, I mean going back to an original question of how do you interact with your warehouse? It's best be able to query off of it and actually build that data tech stack. In an ideal state in the future … as you can imagine, it takes a lot of resources and a lot of knowledge to build that whole thing and to run transformations and manipulate data. And if there's essentially a way to pipe in source data into a certain spot, to do the analytical piece, that would cut out a lot of that middle ground. So essentially you go straight to analytics and you don't need that OLAP layer basically.

Fantastic graphic courtesy of https://www.itransition.com/ — Winson’s basically saying ideally you could shortcut from warehouse to reporting and analytics without any transformation work.


What are the biggest mistakes you see when starting to kick off a strategy and operations function?

Nowadays people are getting more technical on the data side of things, so they're actually building the capabilities of…okay, I need to put my data in here, I have to combine stuff with my product data to my sales and customer data. So they know the warehouse and they know about SQL editors and BI tools.

But the one thing I've found that still is a struggle, whether it's tech debt or legacy company organizations, is the fact all these different go-to-market functions are still pretty much silos. There's sales ops, marketing ops, customer ops, maybe some sales development ops, and everything is very, very much siloed.

Imagine if when the data comes in from marketing and then goes to sales, there's some bad data, so honestly garbage in and garbage out. So that person that's trying to help with customer retention and expansion is just working off of this really bad data and there's no way to go about it. The best way there is to essentially horizontalize everything across the board.


You’ve done dozens of vendor calls in the last few years. what stands out?

To determine the headcount you need or the expectations and goals to hit the revenue target from the board of directors, you have to run a bunch of queries, get a bunch of exports, dump it into the model, have the model update, and you have to populate that across the entire spreadsheet.

In something like Sigma Computing, for example, they just refresh all the queries that pump into the spreadsheet and just give you the updated information to make decisions on your model in your headcount planning or anything else you're trying to do, so these are efficiency gains from different places. And there are different tools I'm also passionate about these days like Polytomic, which combines all the ETL things with a better business model.

If you had an engineering team or a ton of data analysts at your disposal, how would you use that manpower?

Yeah, if I were to think about it, it's like what is the internal platform that’ll enable the revenue engine that’s customer-facing and gets me down to my data and analytics layer? I mean, obviously, there are certain companies doing this now like Apollo, for example, sort of.

How do I get accurate enrichment data for my leads and then be able to have marketing operations say “Okay, we developed some sequences and campaigns. How do we automate that and then have that data tie into opportunity or whatever information in the CRM” and obviously in that CRM there are sales processes or whatever else that work properly to calculate and all that stuff and help salespeople manage that information properly to facilitate the deal.

And once you get to the end, it becomes part of the customer post-sale side of things. So you need a workspace for those folks to retain customers and expand, and ultimately all of that feeds into a data layer of information that needs to be analyzed and used on the day-to-day so that you can help efficiently run different activities and tasks or improve conversion.

This all has to feed into the analytics that comes out of it. That's where the data comes in and says. “Okay. This is happening with this set of customers for this segment…we've realized that this AE or these sets of AE’s are converting these lead sources or these personas they lead to lower retention or lower expansion”

That information across the customer buying journey is super helpful, so having an engine from the beginning of when someone's informed or aware of a product all the way until not just signing the contract, but when they're using a product and have it against the touch points of retaining having all that data in a very seamless data model would be incredibly insanely valuable. And that's obviously where all the data problems are essentially.

One other thing I can think of is automation and workflows across platforms… integrations have to work really well. For anything, it’s just like do you want a unified data model or one that sort of speaks together properly? As I said, if there’s some kind of way to merge data with an OLAP structure and data transformations that’d be incredibly helpful.


Yeah okay, so at the core of all this is unified data?

Yeah, it could be … whether it's a warehouse but nowadays, are there ways to get straight to the sources? Do some computing and then get the information you need rather than all of these warehouse, SQL transformations, combining tables, and you know trying to manually combine data models.


Yeah, totally. Great segueway. You're setting me up very well for a pitch in this article.


Haha yeah.


Second to last question, favorite thought leaders? wWo do you think has really interesting lines of thought about the space and how to grow a startup efficiently?

• Tomasz Tunguz — He’s on the board of Hex, Monte Carlo, and MotherDuck. He basically talks a lot about really, really deep data stuff whether it's warehousing, no SQL versus SQL, or non-relational versus relational.

• Mallik Mahalingam — He's also deep into this space about data warehousing, data lakes, etc. and he knows his stuff.

• Evan Dunn — He created a podcast with his current company Syncari trying to facilitate a lot of conversation around data and how to leverage information across RevOps. He's been interviewing tons of people in his podcast.

Awesome. So how can people keep up with you and what is sort of next on your roadmap or any calls to action here?

Yeah, I'm actually thinking of sharing my own rev ops and data content! So following me on Linkedin would be best. I've spoken to a lot of different RevOps leaders, data leaders, and people in this space who are starting to think “Okay, we have to run a revenue engine really smoothly, which means a lot of things.” The two skill sets used to be very, very much separated, so it used to be strictly systems and configurations (traditional sales ops), and the other was “here's my consulting and banking background” so crunching a bunch of numbers, finding some recommendations, crafting some initiatives, and running with it. But now you need stuff in between — automation and systems talking to each other, so you have to PM that whole system and figure out what the user experience is not from just an AE’s perspective but also the customers’.

Amazing. Great chatting with you as always Winson! He’s been a fantastic advisor and sounding board over the last couple of months — would definitely recommend any early-stage tech founders and data <> growth enthusiasts give him a follow.

If you’re wondering what we’re building here at Cue, you can keep up to date by filling out this form for early product launch notifications and subscriber-only content! We may or may not be launching something involving your cross-system data next week…