To succeed at Product-led growth you need to create meaningful experiences for your end user. To do so you need to quickly operationalize your product data. Chris Im, Marketing Operations Manager at HelpScout covers what someone in MOPs can expect when joining their first PLG company, how to build better data architecture, why Salesforce should not be your single source of truth, and why marketing needs to earn a seat at the product development table.
Chris Im is the Marketing Operations Manager at HelpScout. He has over 10 years experience in marketing and marketing and sales operations in SMB SaaS organizations.
What advice would you give someone looking to join a PLG company who doesn’t have previous experience in PLG?
The first thing is to stay curious and stay humble. You may have built a tech stack and run the campaigns a certain way in your past, but you need to stay flexible. Ask around. You'll find that there are people out there who are smarter than you. You can always learn from them. Practically, that's the first piece of advice that I would give.
Secondly, I've noticed it's really about how much data you can capture, and not just how much data you can capture, but how fast you can do something with it. Taking a little bit of a step back, my experience in terms of building tech stacks for go-to-market teams for marketing, sales teams, and also CS teams as well, I'm mainly using Salesforce, HubSpot, Marketo foundation with all these other tools that are stacked up on top of that. What was really enlightening for me when I joined HelpScout was that there was another way to build your data architecture. Maybe that insight actually meant it was just me being ignorant or naive coming into HelpScout.
Building data architecture a different way wasn’t really ever discussed at the companies that I've worked for in the past. Maybe it was because it was just a smaller team. I’m not sure exactly why it was never discussed this way, but when I came to HelpScout, there was a lot of talk about how there's a better way to build your data stack, be able to capture your data, and then do something with it quickly. The better you can do that as a company, I think the more successful your PLG motion will be.
How does HelpScout Build Their Data Architecture?
The raw data that we get out of our product gets pushed into a data warehouse, which for us is BigQuery. It gets transformed via dbt, and then that's where the magic happens. It gets pushed either to our BI platform, which is Looker, or through a reverse ETL tool. It gets pushed back into our traditional products of Salesforce and HubSpot. I'm sure there's many other tools out there
Our industry has a lot of disparate tools and data silos everywhere. You’ve got HubSpot, you’ve got all these other different tools that are out there. You put it all into one place and then it gets standardized so that you can do things with that data.
In my previous roles, I tried to use Salesforce as my BigQuery, technically my data warehouse, then I’d use Salesforce workflows and process builders to be the dbt tool equivalent. It would basically transform the data to standardize it. Then, using pick list values and different things like that, it can be pushed out and be used by other tools that connect to Salesforce like HubSpot or Marketo. You can do email marketing or send different things about the product, but I had trouble in the past to build that because it’s not really scalable to build your data architecture like that, even though I think a lot of people do it that way.It's good for like one off little things, but to build your foundation like that, you're not gonna get anywhere.
The reason why I talk about this with PLG is because ultimately you're trying to get what people are doing in your product and the data that your own product is capturing into some other tool that does different things. Whether it's a webinar tool, a marketing automation tool, a CRM tool, or a sales automation tool like SalesLoft or Outreach, you’re trying to connect that gap. Don't use Salesforce as that gap. It's not intended to be used that way.
I saw the way they built it at HelpScout is just so fast. You’re able to see the new product feature that was just launched and who actually enabled it based upon this criteria, you can get that subset of people pushed quickly on a daily basis into your marketing animation platform. Then you can put them into a nurturing campaign to let them know they didn’t enable a different feature and ask, “Why not enable this feature?”
At HelpScout, we bring in all the data that comes in from our HelpScout platform into BigQuery before it gets pushed back into our systems. It's actually extremely fast. Before HelpScout, I would call my engineering team or ask them, “Hey, can we get a list of XYZ with this certain criteria?” It would just be a large CSV dump. Then you would have to put that into a list inside HubSpot and match it against your database and then send it. It's so clunky to do it that way, and it was extremely slow. The faster that you can operationalize the data that's coming from your product and get it into your marketing animation platform or your CRM or whatever other tool that you're using, that’s where you will win.
How is HelpScout organized?
We have a typical CS team, marketing team and sales team. Our team is organized with a revenue operations team that sits under finance. They are kind of a third party organization. Under that umbrella is marketing operations, sales operations, finance ops, and a data team. The data team manages what I’ve been talking about.
Then on the CS side, if there’s a feature that was just released that you want to train your users on (and identify who didn't enable that feature) the CS team comes to me. I communicate that new feature request or the data request to the data team. They would then build that into the system. We developed a process through Census, so after the data team builds the request, I can easily connect it with HubSpot. That's when I would build a campaign in the workflow for that new feature.
From a process perspective, we do have a ticketing system that the data team manages in terms of requests. Because we sit under the same umbrella and the same team, we usually get to get things a lot quicker. We're able to speak the same language and we all have the same manager.
How to determine your single source of truth?
Salesforce is no longer the single source of truth. It’s not one service or one software or one platform. The mentality and the mindset that needs to change is that the single source of truth is how to make your whole data stack run so that you can quickly optimize it. Sales doesn't need to see everything. CS doesn't need to see everything. Marketing doesn't need to see everything. You just need to be flexible enough where they can get the right data and for it to be standardized.
The beauty of having everything standardized in a data warehouse or BigQuery is that the data is being standardized in one solution, and then it's being pushed selectively to the different places where people are working. Those destinations may be Salesforce or HubSpot or even Looker. For us, Looker doesn't have everything. it's not tracking clicks on the website. That data is not currently in Looker because it's not needed there in Looker. You can just go to Google analytics and get that information. The fact is that the data could potentially be in Looker if you wanted to be there, because our data is centralized in one place.
Your single source of truth is not Salesforce. It's not one product. It's how you build your full data stack and your data architecture so you can get that standardized data pushed back into those systems.
What about the future of PLG worries and excites you as a MOPs person?
I can start by asking “Why do we like to go into the product?”o one likes being sold to or marketed to. When people point out marketing in a negative way, it’s because they feel like they’re being lied to. One thing that's exciting about PLG is that the product can't lie. Either you have the functionality or you don't. What you see is what you get. The challenge is then how do you match your brand with your product and that product experience. Can you really practice what you preach? If your brand is all about innovation and the latest and greatest, then that needs to be the differentiator between your competitors. If somebody logs into your product and there's zero innovation, then there’s a mismatch between expectations and results. That’s exciting and that’s scary.
Marketing now has a bigger play right in the product development cycle. In the past it’s been about advertising and making pretty pictures and things like that. It’s now become part of a sales cycle with the advent of online business. We’re now creeping our way into the product life cycle. What's exciting is whether marketing gain that seat not just at the revenue table, but can we gain that seat at the product development table as well? Can we play a role with how a product is developed for both users and even how backend structures are created? Can we use that as a marketing tool?
For example, we can build a system to filter out spam and capture how people sign up for products. Can we capture logins? Can we capture login durations and seats and what add-ons they're able to add on to their subscription?
This seat that I talk about at the product development table is a seat not easily gained. Marketing really needs to gain that respect. Marketing has become a little bit lazy in the way that we market. It's so easy to just push a button and then send out a half baked ad to just all your ICP. Maybe we've gotten a little bit lazy, so we’ve got to start thinking a little bit more critically.
The problem these days is that there's just so much tech out there, specifically SaaS. No one really reads Scott Brinker’s “MarTech Landscape” anymore because there’s just so much tech out there now, but that's our current landscape.
A lot of tools look the same. How can you really differentiate yourself? That's why PLG is so exciting. It's really about the first impression, which is just so important. You can't fake it anymore by throwing a lot of money at it. You can pretty up your website with pretty pictures and make it look nice and even have great content. You can even manipulate the sales process. Let’s say somebody requested a demo and was given a really good demo instance, but they’re not allowed to actually trial the product. They buy it. Then the moment they log on for the first time to use the real product, it’s a completely different experience than what they experienced during the demo. The product cannot lie. That’s why PLG is so enticing because what you see is exactly what you get, the product can’t lie.