Table of Contents
- • No silver bullets: Building the analytics flywheel
- • Identity Crisis: Navigating the Modern Data Organization
- • Scaling Knowledge > Scaling Bodies: Why dbt Labs is making the bet on a data literate organization
- • Down with 'data science'
- • Refactor your hiring process: a framework
- • Beyond the Box: Stop relying on your Black co-worker to help you build a diverse team
- • To All The Data Managers We've Loved Before
- • From Diverse "Humans of Data" to Data Dream "Teams"
- • From 100 spreadsheets to 100 data analysts: the story of dbt at Slido
- • New Data Role on the Block: Revenue Analytics
- • Data Paradox of the Growth-Stage Startup
- • Share. Empower. Repeat. Come learn about how to become a Meetup Organizer!
- • Keynote: How big is this wave?
- • Analytics Engineering Everywhere: Why in the Next Five Years Every Organization Will Adopt Analytics Engineering
- • The Future of Analytics is Polyglot
- • The modern data experience
- • Don't hire a data engineer...yet
- • Keynote: The Metrics System
- • This is just the beginning
- • The Future of Data Analytics
- • Coalesce After Party with Catalog & Cocktails
- • The Operational Data Warehouse: Reverse ETL, CDPs, and the future of data activation
- • Built It Once & Build It Right: Prototyping for Data Teams
- • Inclusive Design and dbt
- • Analytics Engineering for storytellers
- • When to ask for help: Modern advice for working with consultants in data and analytics
- • Smaller Black Boxes: Towards Modular Data Products
- • Optimizing query run time with materialization schedules
- • How dbt Enables Systems Engineering in Analytics
- • Operationalizing Column-Name Contracts with dbtplyr
- • Building On Top of dbt: Managing External Dependencies
- • Data as Engineering
- • Automating Ambiguity: Managing dynamic source data using dbt macros
- • Building a metadata ecosystem with dbt
- • Modeling event data at scale
- • Introducing the activity schema: data modeling with a single table
- • dbt in a data mesh world
- • Sharing the knowledge - joining dbt and "the Business" using Tāngata
- • Eat the data you have: Tracking core events in a cookieless world
- • Getting Meta About Metadata: Building Trustworthy Data Products Backed by dbt
- • Batch to Streaming in One Easy Step
- • dbt 101: Stories from real-life data practitioners + a live look at dbt
- • The Modern Data Stack: How Fivetran Operationalizes Data Transformations
- • Implementing and scaling dbt Core without engineers
- • dbt Core v1.0 Reveal ✨
- • Data Analytics in a Snowflake world
- • Firebolt Deep Dive - Next generation performance with dbt
- • The Endpoints are the Beginning: Using the dbt Cloud API to build a culture of data awareness
- • dbt, Notebooks and the modern data experience
- • You don’t need another database: A conversation with Reynold Xin (Databricks) and Drew Banin (dbt Labs)
- • Git for the rest of us
- • How to build a mature dbt project from scratch
- • Tailoring dbt's incremental_strategy to Artsy's data needs
- • Observability within dbt
- • The Call is Coming from Inside the Warehouse: Surviving Schema Changes with Automation
- • So You Think You Can DAG: Supporting data scientists with dbt packages
- • How to Prepare Data for a Product Analytics Platform
- • dbt for Financial Services: How to boost returns on your SQL pipelines using dbt, Databricks, and Delta Lake
- • Stay Calm and Query on: Root Cause Analysis for Your Data Pipelines
- • Upskilling from an Insights Analyst to an Analytics Engineer
- • Building an Open Source Data Stack
- • Trials and Tribulations of Incremental Models
New Data Role on the Block: Revenue Analytics
In 2019, the data community realized there an emerging role called Analytics Engineer.
In 2020, we started to talk about Operational Analytics.
In 2021, I am building out a data function called Revenue Analytics.
Browse this talk’s Slack archives #
The day-of-talk conversation is archived here in dbt Community Slack.
Not a member of the dbt Community yet? You can join here to view the Coalesce chat archives.
Full transcript #
[00:00:00] George Sass: Hello, and thank you for joining us for Day 5 of Coalesce. My name is George Sass and I’m a senior software engineer at dbt Labs. Today, it’s my pleasure to be hosting Celina Wong who’ll be talking to us about the New Data Role on the Block: Revenue Analytics. Celina is currently the director of analytics at TULA Skincare and has an impressive background in both data and finance. So she might just be the best person to talk to us about this emerging role at the intersection of those two spaces.
All chat conversation will be taking place in the coalesce-revenue-analytics channel and the dbt Community Slack. So if you’re not in that channel already pop on over and join us.
As a reminder, the Slack is a place for you, the audience. We encourage you to ask other attendees questions, make comments, or react to any point in the channel. Really, have at it. [00:01:00] I’m personally hoping for great memes and even better, pop music call-outs. So post those Spotify wrap-ups if you’ve still got them saved to your desktop.
After the session, Celina will be available in Slack to answer your questions. Feel free to save those for after the session or just post them in Slack as they come up. Let’s get started. Over to you, Celina.
[00:01:29] Vince Cimo: Celina, just a heads up. You’re muted. You’re totally good.
[00:01:34] Celina Wong: Let’s redo this party intro. Thank you so much, George. And who else loved the intro music? Hello. Hello and hello to all my purple party people. Welcome. What did you just sign up for? Settle in, cozy up. You’re about to take a trip down a musical memory lane with me.
What does data have to do with music? [00:02:00] You’re about to find out. So hang on to the edge of your seat. First and foremost, I am Celina Wong, currently director of analytics at TULA. We are a skincare startup focused on creating products that are from pre and probiotic extracts. We’re a hundred percent remote. LinkedIn last year named us one of the top 50 rising startups all to say we are hiring. A bit more about myself in revenue analytics.
I started my career in corporate finance over a decade and eventually made the transition into analytics in the startup world. And I believe that positioning analytics and revenue forecasting together is quite powerful for not only the business but for the analytics team.
[00:02:48] What’s happened in the last 2 years? #
[00:02:48] Celina Wong: So let’s talk about what’s happened in the last two years. Does anyone else feel like dbt is that best friend that puts a phrase to something that you quite can’t explain, [00:03:00] but you needed a term for it. In 2019, if you were in dbt Slack, or if you scroll back far enough, you might’ve heard the term, and now we all know it, analytics engineering buzzing around in Slack. And what we didn’t realize was that we were identifying a new role, right? We were doing something that we didn’t know how to name it. And now it’s one of the hottest jobs on the market. And then in 2020, we started to hear the key word or the term operational analytics.
And what did it mean? It was talking about tools and process and putting data to work. Well, over the last few days of Coalesce, we’ve been asking ourselves, what is next? What is the future of analytics? What tools or roles are we not talking about yet? We’ve come to almost the end of 2021. And I would say, let’s start [00:04:00] talking about revenue analytics.
Let’s start using data to inform decisions. It is either putting a term to something that you are doing already and didn’t know how to name to it, or you will be doing after this talk.
So remember when I mentioned that you were about to take a trip down a musical memory lane, buckle up and here we go. If you didn’t know what the title of this talk was inspired by, then let me give you a big hint as to who, one of these boy bands are, drum roll, New Kids on the Block. You’re very welcome.
Here’s a flashback to the eighties and early nineties. And if you know these bands, please give them a shoutout in Slack as well. There are no prizes except for pride in recognizing all these music icons throughout the presentation. [00:05:00] Didn’t realize you were coming to music trivia, did you? Well, have fun and enjoy these music references as we embark on the journey of how we created a new data role called revenue analytics at TULA.
How did we get started? What sparked the beginning of this new role? Our VP of e-commerce, and for those of you who are not familiar with the word e-commerce, this role runs our website acquisition and retention marketing was responsible at the time for forecasting, our DTC, which stands for direct to consumer business.
And it was painfully manual to forecast and he thought there’s gotta be a better way. It was exposed to manual error, human bias, and a very long turnaround time. It was limited to month over month, quarter over quarter, year over year trends. And we know that’s not indicative of the business moving forward, especially at a startup.
[00:06:00] And as our business scaled, our VP of finance was also nervous about the methodology at the time and knew that we needed a better way to forward and predict the business.
So what happened next? Our CEO had mentioned that she had a finance partner at her previous company where they used a cohort forecast model to help predict revenue. And so we worked together to build a model for TULA. So cheers. So the power of connection, community, and helping each other learn something new.
So you might be sitting there thinking, what is this cohort forecast? You may be doing this already today and if not, listen up, it may apply. We built a linear regression model, which ingests historical data by covert and predicts repeat orders for future months. And we layered in additional levers, [00:07:00] such as expected marketing trend, CAC, which stands for cost to acquire a customer. CPO, which stands for cost per order, AOV, which is average order value, etc. I know. Quite a bit of lingo all to get to a revenue forecast.
That’s great in all. So how do we fold this into the organization? This group of producers, yes, yes, I know one of them is not the same as the others, I’d like to argue that the answer is skillset. You might be sitting there saying no, Celina, it’s fame. We could debate that in Slack later.
But in any case, we knew that we needed the right skillsets to maintain and improve and understand this model. Forecasting has traditionally sat under the finance team. But for this particular model, we knew it wouldn’t fit under the finance [00:08:00] team.
[00:08:01] Implementing revenue analytics #
[00:08:01] Celina Wong: So, we formed a new band under the analytics team called revenue analytics. Great band name. It is a new dedicated analytics function and resource. And what we saw was company-wide success and the crowd went wild for these four key results. One, accuracy. We saw a 25% accuracy improvement. Two, speed. We saw faster turnaround time from two weeks to two days, which means more time for feedback and strategic discussions versus just the mechanics.
Numbers three and four. We love to hear this, trusted strategic partner and we gained a seat at the table with leadership to discuss the customer trends and insights that we were seeing and how that drives the model assumptions forecast and what it meant for the [00:09:00] business.
But of course, company-wide success is not enough, we need to talk about global success. Now that’s great in all for our analytics team. Great. You can sit at the table. Fantastic. But what does this mean for you and your team and your organization? What does this mean for the data community and data organizations across the globe? Shoutout to the BTS army. Yes, you have permission to dance.
In these last few days, we’ve discussed the importance of topics like self-service to operationalizing data. Part of this means it’s time to cross the bridge from being just a data-wrangling analyst to strategists. And you might be sitting there going that’s great, but how? Well, let’s think about it.
[00:10:00] What do most organizations worry about? What’s keeping your leaders, your CEOs up at night? You may be forming some thoughts out there. Let me tell you a keyword is revenue. Put yourself and your team in a position to help the organization feel more confident about where revenue will land and what are the levers together.
And that means when companies have this revenue analytics role, this is the overall impact for data orgs as well. One, you are not just a data wrangler. And I repeat after me. You are not just a data wrangler. You can be a business thought partner. You can and should be informing business decision-making. And last but not least, more proactive driving versus reactive writing on requests. Having a data [00:11:00] role focused on mission, critical challenges, such as predictable revenue, offers the opportunity for both the business and the data team to better leverage the organization’s dataset and data talent.
[00:11:21] Why wasn’t revenue analytics always a thing? #
[00:11:21] Celina Wong: That’s amazing. But you might also be wondering companies out there have been worried about revenue for a long time. So why wasn’t this always a thing? Great question, my music trivia champs out there.
We thought we were covered. We thought we were covered and assumed that having a forecast built by the inventory demand planning team and another one, which was a roll-up where each marketing channel lead. So you may have folks who are leading your acquisition area. You have folks who are leading your retention efforts, [00:12:00] rolling up their forecasts to what they believe next year will look like.
And we thought that would be enough to feed into an overall finance team’s budgeting and team target setting process. However, these were all a manual process highly subjected to human error and bias. And then we realized there was a gap and the gap involved levers, like letting a model predict returning orders instead of fitting a forecast to someone’s expected mix of returns of returning customers.
What is that based on? And so we re-imagined. Reimagined how to strengthen the process by adding a revenue analytics role, to own the predictive revenue forecasting process from the lens of customer cohort data, to help triangulate a more accurate overall budget and team target.[00:13:00]
[00:13:01] Why is revenue analytics emerging now? #
[00:13:01] Celina Wong: And so why is this role emerging or why is it emerging now? For the business and from the business’s perspective, we’re now in an environment where companies recognize, and we’ve been talking about it the last few days here at Coalesce, the need for data teams. And we’ve seen the exponential growth of data tools such as dbt that help us ingest, transform and automate the, what is happening part of the data maturity curve. It is time to leverage the amount of data we’ve collected and the data talent that we have to get to what will happen. Using data to predict the business is not optional anymore. And for us as data professionals, we need development. We need growth. We need new challenges such as the revenue analytics role where you can leverage the skills you’ve built as an analyst or an analytics engineer and apply it [00:14:00] to strategy, apply it to what will happen next.
So if you’re sitting out there and still holding onto the edge of your seats and you’re fired up and excited to take the next steps, what does it mean for you? There are two perspectives. One, if you’re a data leader out there or a leader, who’s thinking I need to hire for this role, what skills should I be looking for?
[00:14:29] Skills needed for revenue analytics #
[00:14:29] Celina Wong: Here are a couple of tips or bullet points that I would share with you when looking for a candidate for this role. One is quite obvious, but stating Python, SQL, Excel skills. Someone that has a solid business acumen. Someone that may have an understanding of how a P&L works, P and L stands for a profit and loss statement. And a candidate who may have finance experience or background. It may be someone who’s looking to transition from finance to under [00:15:00] analytics, and always, don’t forget to see the potential in someone. Just because they may not have exactly every single bullet point does not mean they don’t have potential. Now if you’re an individual out there and you’re really excited and you’re thinking to yourself, I need to position myself for this role. Here are a couple of pieces of advice that I can offer you for how you could get into a role like this one. Go learn how a P&L works. What sales or marketing channels are driving your company’s revenue? What’s driving its costs? Go meet with the finance team of your organization. They may also be called the FP&A team which stands for financial planning and analysis. Ask them what their pain points are.
Are you hearing about, we need to dig into more data, but we can’t get there. How do I build this relationship to learn more about how I can help drive business impact [00:16:00] and bring what you learned back to your data leader. Explain what you’ve learned and where you think the data team can jump in and help with revenue analytics.
And then turn around and go build it. Build this model, show the variance to actuals, go prove the accuracy and the reliability of this method.
Many community members out there are already beginning to invest in the revenue analytics role, including dbt Labs. If you caught Ric Louie talk on Day 1, she is building out this function in 2022 as well. And so I leave you with this challenge: Is your company next?[00:17:00]
Thank you so much for your time today. I wanted to give a shoutout to my TULA team in the audience, especially to Tyler who’s helping me shape revenue analytics at TULA. A special Thank you to Mary, George, Winnie, and Jillian for supporting this talk and a sincere thank you to all the folks at dbt Labs for making data fun while welcoming people coming from all backgrounds.
[00:17:29] Q & A #
[00:17:29] George Sass: And thank you Celina for a truly amazing talk. The chat has been going absolutely wild. I cannot overstate how often several people are typing came up. And they’ve got a lot of fantastic questions for you. We do have some time to answer them right now. I’m going to start with one of mine though.
What pop single do you recommend everybody go listen to after this talk?
[00:17:54] Celina Wong: George, I’ve been thinking about this and the only thing I’ve been able to come up with [00:18:00] was NSYNC’s Bye, because of the end of this talk. We’re not at the end, but I can’t help but think about the dance moves, and you all know what I’m thinking about out there. So that’s why that’s top of mind for me.
[00:18:12] George Sass: Fantastic. All right onto other less populated questions. So Janessa is wondering, I would be interested in hearing Celina say more about how they measured the improvements in speed and accuracy. Metrics can be hard.
[00:18:29] Celina Wong: Yes. And that is a fantastic question because I think in data we’re often trying to prove, right, what is improvement? What is our impact to the business? I think from the speed perspective, I’m talking about going from two weeks to two days, it was a combination of having straddling your VP of e-commerce to come up with a forecast, right? Someone who already has plenty of other priorities on their plate and then relying on them to come up with a [00:19:00] forecast versus now having a dedicated resource on the analytics team who has a model that they’re working with, and the ability to plug in data and turn it around within two days so that the team here can also go back and give feedback, right? Because we all know the first time around with any data model is not going to be perfect. If it was, please let me know where you found it. But I think the importance of removing the mechanics of things and allowing for more time for discussion.
And then in terms of measuring the accuracy improvement, actually, Tyler, my team member here in the audience, has actually gone back to the model and rerun the model several times. And so in terms of measuring it against how we did it before versus now we’ve actually gone back and looked at prior versions of the forecast versus with the model, what it looks like now to provide that increase of improvement not only to prove it to ourselves, like how reliable is this model that we’re showing the rest of the company, [00:20:00] but also to show to leadership. It’s always important, right? I think it’s a theme that’s come up across Coalesce is how do we showcase the work of analytics? And I think this was a huge win for us at TULA.
[00:20:12] George Sass: So we’ve got another one from a Winnie who is asking, is it easier to teach analytics engineering to finance people or vice versa? How would you go about thinking about this?
[00:20:24] Celina Wong: Great question, Winnie. I actually think it’s easier to teach finance to analytics engineers and the reason being, I think that when you have a engineering or analytical mindset already, but you’re not familiar with, just what is the lingo? How do I think about the math behind a financial model? It’s a bit easier to teach where I would say it’s a shorter period of time to learn than to have to teach someone in finance, all the technical skills of building up Python and SQL.
[00:21:00] Somebody out there may be fantastic. And so they might argue, vice-versa but this is the path that I’ve seen so far.
[00:21:07] George Sass: Makes sense. So Josh Devlin asks, I’d love to understand at a high level, how your forecast are generated. Is it ML? Formulaic?
[00:21:19] Celina Wong: Great question too. It’s actually a combination of we’re running our data through a Python model that is then spitting out what the expected return orders are gonna look based on prior cohort behavior. And then we’re also formulaically adding in the layers of, for example, we know our marketing spend is going to change. We have expectations for what we think traffic will look like in the future months. For those of you in e-commerce, we’ve seen the ride of 2020 and 2021, right? So when it comes to ML or when it comes to, Hey, I’m going to rely on a machine to tell me what will happen, there are times you have to plug into the model and say, This is [00:22:00] likely not going to happen because historically we’ve never seen an event like this before as we saw in 2020. So the answer is it’s a combination and it will continue to be a combination, which is why I think the business acumen piece and understanding your P&L piece of the role is highly important because like with any analysis, with ML, you might spit out, you will do X amount in business. But do you think that’s realistic and that’s where an acumen business acumen understanding of P&L and discussions with your leadership team come into play.
[00:22:33] George Sass: We’ve got another one from Mila who’s asking, how would you encourage a budding analytics practitioner to build the soft skills needed to complement or support technical skills?
[00:22:45] Celina Wong: My number one advice to a budding analysts looking to gain more soft skills is start with those one-on-one conversations. It’s really important. And I know at times we sit there and go, I have so much work to do. I have to [00:23:00] put my head down, do my work. But I can’t emphasize enough the amount of times that I had a one-on-one chat or a coffee chat with folks inside the company and outside that have helped shape my perspective.
A lot more than just me, focused on just my project, my work, delivering it, including my laptop. Because there’s only one of you, but there’s so many things to learn and perspectives. And I also think those one-on-one conversations help you build more confidence when it comes to the soft skills of perhaps talking to someone that may feel like they’re a lot more senior.
But don’t forget, they’re also human. And I think that’s where you start to feel like you’re forming more of those soft skills, because once you start those one-on-one conversations, you also build this, I would say it’s like my trust network, right? They’re the ones encouraging you to get up on stage. Like I did today to present in front of all.
[00:23:55] George Sass: Thank you. So we’ve got another one from Janessa who side note is [00:24:00] Darren you to do the dance moves? She’s asking what are the signs of potential that you look for? Curiosity, interest, a certain aptitude in something?
[00:24:10] Celina Wong: Yeah. I think many data leaders out there feel this way where when we look at a candidate and we’re searching for potential. I look for the eagerness and hunger to learn. And I think that speaks to like curiosity and whatnot. But I think that the enthusiasm and passion for wanting to get into something or wanting to gain a particular skillset has proved to me time and time again, someone’s ability to be successful in role, even if they don’t check all the boxes or qualifications for that job.
[00:24:41] George Sass: So someone who’s hype.
[00:24:45] Celina Wong: Yeah, very hype. Just like this talk, hopefully.
[00:24:49] George Sass: I think it’s a hype talk. We’ve got another question from Alexis who’s asking, how do you see your revenue analytics fitting into the overall hiring plan for a data team? [00:25:00]
[00:25:01] Celina Wong: Yeah. From an overall hiring plan I think when you let’s say you’re starting out and setting up your data team and your data hire number one out there, I don’t think revenue analytics is going to be your first hire. I’m not going to go and argue that because I think there is a higher need for an analytics engineer or data analysts who could be more of a generalist to cover all the areas that you are covering at the time. But I do think that revenue analytics may be your second or third hire because this individual or this role is going to give you that seat at the table.
A lot of us argue about how do I get more resources on my team? And oftentimes we’re talking about the ROI of your team, right? And I think that revenue analytics is where you can prove yourself or prove the ROI of a data team. And so while I don’t think it’s going to be your first hire, I do think that it will be second or [00:26:00] third hire makes sense.
[00:26:03] George Sass: Kevin’s asking about something that you touched on a little earlier which is, has COVID made revenue harder to forecast as the world changes significant?
[00:26:15] Celina Wong: Kevin, great question. The short answer is Y E S yes. I don’t, I have, just mind boggled by how it’s changed revenue analytics and the ability to predict, and it’s been a wild ride, but I still think that with our model actually internally at TULA, we’ve done a good job compared to if we had to just let the process keep going the way it did before. But yeah, to summarize, yes, it has thrown a, I was going to say a wrench,but I feel like it was throwing multiple wrenches at us at any given time.
[00:26:52] George Sass: Yeah. As it did at everyone, I’m pretty sure. So we’ve got a question from Sam Harding. Sam is asking, [00:27:00] something I haven’t heard a lot of people talking about is using data to improve supply chain, such as inventory management, warehousing distribution. Is that because we’ve been doing it for a while now, so it’s not a hot topic? Is it being overlooked or am I just not talking to the right people?
[00:27:18] Celina Wong: . Really great question. And my answer is it’s being overlooked. It’s one of those things. Similar to how forecasting has been done for a long time. It’s been around. But I think the overlay of data and analytics and taking someone who is in the data organization and putting them on supply chain or inventory forecasting, for example hasn’t come across yet.
My view when thinking about the future roles that we see out there, I think there’s been plenty of conversations about, are we going to see sub segments of roles, as analytics matures, and I’ve already seen it in certain companies where, we’ve got revenue analytics, but I could [00:28:00] see supply chain analytics also being a function in the organization.
If that is an area that has a lot of just strap capital because of poor planning. And so if you think about that’s a heavy cost to your organization where they could have better use that cash somewhere else. And so that’s impact right there and that, so I believe it’s overlooked. I’m actually interested in this as well.
[00:28:26] George Sass: What other future roles do you think might come out of the diversifying analytics engineering?
[00:28:32] Celina Wong: I think that a couple of other talks at Coalesce have spoken about it already. But I do think that when it comes to the data analyst role we’ve been generalists for quite a while. Now we cover every department group, and I’ve seen and discussed with many leaders, how we think about centralized versus decentralized. How do we support different groups? And so my team is centralized, but I do think about how the way our teams organized is we have [00:29:00] owners that ladder up to a business partner. So while we’re centralized, you still have a one-to-one business partner and they know who to go to versus having a model where I’m going to go to the data team and I don’t know who’s going to pick it up.
And I think the other advantage is having contexts. When you’ve had that one-on-one relationship, you have more context to jump into requests faster. And so to answer your question I think revenue analytics, supply chain, operation analytics we have someone who is focused on digital marketing analytics, and I think that’s a role we’ve all seen and has been around for a long time. And then there’s also thinking about, of course product analytics has also taken off. And then I also have retail analytics where we are starting to see this shift in 2021 already. And I’m curious to see how the macroeconomic changes things as well, but for businesses out there like ours, where we have a direct to consumer website, but we’re also selling [00:30:00] to consumers in brick and mortar, what is that analytics going to shape out to be? And that gets a bit more complicated because it depends if you own your own store location, or if you’re working with the retail partner out there. So TULA’s at Alta. And so we have to work with the data that we receive back. And so that actually has its own subset of analytics as well that we’re going to get into at TULA.
I say it’s dependent on how your company is shaped, but that’s how it’s shaped at TULA and I would imagine at other e-commerce companies as well. And so those were sub segments of the data analyst role that I could foresee shaping up in the future
Last modified on: Apr 19, 2022