July 6, 2017
Ep. #5, WebTorrent: Bringing BitTorrent to the Web
In the latest episode of Demuxed, Matt, Steve and Phil are joined by Feross Aboukhadijeh and John Hiesey for a discussion on WebTorrent.
In this Heavybit Workshop Panel, Peter Chapman is joined by HackerRank CMO Nicolas Draca, Andrea Echstenkamper, Director of Marketing for LaunchDarkly, and Clearbit COO Luke Whiting. The group discusses the tools and strategies for measuring, monitoring and maintaining growth from day one.
When should you start using a CRM? What are the right user metrics to follow? How long before Google Sheets won’t cut it? Hear the answer to these questions and more from a panel of modern marketing experts.
Peter Chapman: Welcome to Heavybit’s workshop panel on early stage growth ops.
I’ve been talking to a bunch of you folks in the audience about building early-stage growth engines. And I realize that, while I have a playbook or the platonic ideal of a sales and marketing machine, I didn’t really know what those first steps looked like.
At what stage do you need to invest in Salesforce? At what stage do you need to invest in an email automation tool? What stuff do you need to buy tooling for, and what stuff can you get away with in Google spreadsheets?
I assembled a team of really crack operators to talk us through their experience building out these growth engines, and I’m super excited to dive into it.
We’ve got Nicolas Draca who runs marketing at HackerRank and was formerly the CMO of Twilio. We’ve got Andrea Echstenkamper who runs marketing at our very own LaunchDarkly, and we’ve got Luke Whiting, the COO of Clearbit, formerly of Heroku.
You guys ready to dive in? I want to start with you and the products you’re selling now. Maybe we could just hear a little bit about what you’re selling and who your audience is.
Nicolas Draca: Yes, and a quick fix, I wasn’t CMO of Twilio, as it’s recorded, I was VP of Marketing at Twilio.
Peter: Oh, thank you.
Nicolas: There was somebody leading marketing. So, today I’ve been here currently for six months. Before, two years at Twilio, five years at LinkedIn. And at HackerRank what we do is we match every developer to the right job.
We have a community play where, as a developer, you can come and learn to code, practice to code, and compete. That’s for the community side of the business.
We do not monetize, the platform is free. On the other side we have a platform, an assessment platform, allowing a recruiter and hiring manager to assess a developer through the hiring process that we sell to recruiter and hiring managers.
Peter: Cool. Andrea?
Andrea Echstenkamper: Great. I’m at LaunchDarkly and what we do is help eliminate the risk from development cycle as well as help you deliver value to customers faster and with more control. And a lot of that is through the magic of feature flagging. We sell primarily to the VP of Engineering and the “Head” of certain departments within engineering and product.
Luke Whiting: I’m at Clearbit. We are a business data provider. We sell primarily to sales and marketing teams who want to get to know their customers better.
You send us an IP address or email address and we can tell you any number of things about those customers from job titles to locations to employee counts for businesses. We can also help you find new customers with our prospecting tools.
Peter: I want to start by talking about metrics. Let’s imagine that you just stepped into an early stage company, there is no marketing team, and your job is to manage growth. What do you start tracking from day one?
Nicolas: Anyone want to start? I had a chance to do that exercise actually recently with a company here.
There is an assumption that there is no metric, and I think that assumption is wrong. You have data. You have something that’s going to help you get somewhere.
What I mean by that is, at worst case, you have your website traffic, you have some concept of signup or conversion depending on what type of call to action you put on your website, and you have a concept of revenue further down, even if you’re not making money. But there is a concept of pipeline and revenue that you can define any way you want, but those concepts do exist.
The first thing to do, I think is, you ask about CRMs and complex systems and take a Google sheet and try to build a concept of cohort on a monthly basis trying to define a set of stages, again, whatever are those stages. You can start with web traffic, sign up.
After, I’m sure we’ll go deeper, but different stages, and start capturing the data that you have and put them in that Google sheet and looking back at the data. Whenever you started your business is when you have the first data, and that might not be perfect, but this will give you an idea about what you have and an idea about conversion rate.
Andrea: That’s exactly how we did it at LaunchDarkly. Start out with this sheet where it’s going to be your source of truth. Maybe you’re pulling it from a lot of different places at the beginning, and then later on when you have systems, you’re still probably pooling them in there so that there’s one place where you have that information that existed at that exact time.
But another layer is to also put in, what marketing campaigns, what influential things happened during that week or that month so that you can see what’s really affecting those individual metrics.
Luke: I think it’s really important, especially in the early stages, that you don’t neglect talking to customers.
I think you build out a funnel, but really, especially in the early days, you should be taking to customers.
It’s the kind of thing that doesn’t necessarily scale. You have a self-serve model, you need to be getting in touch with those customers. Why did they buy? What did they find interesting about your product?
I feel like, from my perspective, you really need to be able to understand those kind of fundamental questions before moving on to the, maybe the more data driven side of things.
Peter: Andrea, Nick, do you guys do qualitative research up front?
Nicolas: Yes, so you do, good catch on this. I would consider it, I’ll put in the data bucket, it’s customer knowledge, customer research. Either you do it manually, depending on the volume you get, like you say, it’s time consuming. But it’s worth it because you cannot assume and you want to understand, try to identify the use case where people have been able to move through that funnel and get somewhere.
Again, you can assume things, but be open to reassess and to really take and listen to customer feedback because at the end of the day, the “customer is king” and they are the one buying.
If you go back to metrics, the metrics I would be tracking, I have a concept of KPI and metrics, I track three KPIs. One will be email opened, meaning in any communication. It looks like a super tactical metric, but it’s just to see for marketing campaign if yes/no, you are engaging with the right people or if yes/no your content is relevant.
I tend to move out of that “email opened” pretty quickly, but I think to get started is a good way to try to get a sense about the quality of the database and how you’re doing from an execution standpoint.
The second one, depending on the type of business you’re in, it’s either SQL, pipeline creation as a count. Or if you’re in the on-demand business, self-serve, I think the concept of a product active.
I’m sure we’ll talk more about funnel, but defining a stage in the funnel which is not “the person’s visited thewebsite, the person didn’t sign up,” that’s not really going to help. But they did some meaningful interaction within your product that you have to decide with yourself, I have to say, of what this is.
So a count of product active by cohort and yeah, I’ll go for two. I’ll go for those two. And if I go for a third one, actually, is the number and I’m going to piggyback on what you said, the number of times you talk to a customer. I think you have to set this as a metric. Is like five, meetings a week, whatever it is.
Luke: It’s like very early stage metric you’re tracking is, how many conversations am I having a week?
Nicolas: Yeah. And this would be a forcing function because then you’re going to say I haven’t done it this week, I have to do it because quickly we do it like for the first two weeks and then forget about it. I really like that.
Luke: I feel like
once you understand the profile of your ideal buyer, then you can start tracking qualified leads.
You have a profile and then it’s what marketing activities are driving qualified leads, like what are the channels or methods that are driving?
Peter: You guys both use this terminology. Nicolas, you called it an SQL. Luke, you’re referring to a qualified lead. What is a qualified lead in this context?
Luke: Someone who’s like most likely to buy your product. I think you can measure that from a buyer profile or persona. Or it’s, if you want to get really detailed, it can be what activities are they performing on your site?
Have they signed up for an account or are they playing around with the free trial? Have they viewed marketing pages? If you have white papers, have they downloaded white papers? I mean, those can all kind of play into that scoring model.
Peter: Cool, we’ll come back to SQL later. I want to talk a little bit about tooling. Nicolas, you said day one totally fine to just dump all this stuff in a Google sheet and Andrea, you echoed that. What are the fundamental pieces of infrastructure that I should start to invest in from the early days?
Andrea: From there what do you start thinking about, eventually you are going to need a CRM. You are going to need email automation, although that becomes marketing automation at some point.
You’re going to want to set up Google Analytics as soon as possible with goals so that you could see what are these sources that people are even spending time on our site even before they get to sign-up.
Is there a problem before there? Even just to take some hypotheses about who you think your audience would be and who’s going to pay attention. And then also at some point you’re going to want to look into Segment.
Peter: What is Segment? How does that fit into your stack?
Luke: Segment… everyone’s looking at me. We’re we’re big fans of Segments at Clearbit. Segment is essentially a marketing data hub.
The idea is that you can feed data from multiple sources into Segments and that can be through one of their integrations where they’re pulling data from Salesforce are pulling data from another marketing tool or it can be completely custom where you’re using one of their APIs to dump in your own analytics data.
The advantage of using Segment is that once you have the data in Segment, you can then fan it out to any other source where it might be useful. So as an example at Clearbit, we dumped all of Clearbit data and a marketing activity data into Segment and then we then push it into our marketing automation tool. We can then send emails based on the data that we have from Segment.
We also dump, all that data eventually flows through segment and into Redshift, our data warehouse. We can then build SQL queries that show us the entire lifecycle of a customer’s relationship with Clearbit all the way from the first page view down through the purchase and closing of an opportunity in Salesforce. That’s a high-level overview of what Segment provides and why we find it very valuable.
Andrea: And then in addition, to bringing all these together, what is this one user doing across all of from Zendesk to production app to Salesforce? There’s also benefits around protection for your app, just security. And then there’s also the fact that there’s less loading time, for SEO purposes.
Nicolas: Just right on this, because it could be also pretty confusing, as of today you can pick from, I don’t know if it’s 2,000, 3,000 pieces of technology in marketing, as of what is the latest number? I’m looking at your end of business.
Luke: That number sounds right.
R- It’s pretty overwhelming.
What is exciting is you can solve anything you want with technology as of today. What is less exciting is, first it’s pricey, two, it’s overwhelming. Because more technology you have more technology you’re going to have to manage.
And then, as we have developer in the room, you’re like okay, I can build it myself. Which is a bad idea. Don’t do that.
I’m going to tell you why, because if you build it yourself it means you’re going to have to maintain it. Which means after, you’re going to hire a marketer who’s not going to be able to use it and it’s a chicken-and-egg the number of hours you’re going to put in coding these things.
So, long story short, you want to start simple. And what you need, as you said, is CRM and email marketing which will become, one day, marketing automation.
You have two paths. I don’t work for any company in that space, by the way. You have a path where you can go HubSpot. Many people start with HubSpot, this is what I’ve seen. Or they start on HubSpot as a free CRM, I’m not sure about the last piece but I think CRM is free as part of the package.
You have to get started somewhere or you go directly Salesforce, and Salesforce acquires products and it’s supposed to be fully integrated. I haven’t been using it but this is what I’ve seen many startups do when they get started.
There is a price to it, just to be clear, but it is worth the investment because if you don’t do it then you’re going to either build your own thing or you going to spend cycle trying to understand what’s going on and you will not know what’s going on.
So there is a little bit of a premium. But you need to start with that foundation of email marketing CRM for the thing to work and to equip your team to see what is going on.
Peter: So you absolutely need email marketing and you need CRM?
Peter: And Salesforce plus Pardot or HubSpot plus HubSpot fill both those roles.
Nicolas: This is what I’ve seen but I don’t know what you–
Luke: I think there’s pre-phase where your CRM is your inbox and you’re doing analytics in Google Sheets, which is totally fine. And I think is where you should start. I’m of the opinion on the CRM side that all roads eventually lead to Salesforce and so the sooner you can bite that bullet and invest in Salesforce, the happier you’ll be down the road.
Most people that I’ve seen who’ve started elsewhere end up moving to Salesforce at some point eventually. And you’ll just save yourself the headache of switching if you just start there.
Peter: Andrea, I see you nodding and smiling.
Andrea: Agreed, yeah. It’s just kind of inevitable.
Peter: So at what point do we bite the bullet? When do you have to move from my inbox to, I need to spend money on Salesforce?
Nicolas: I think one of the drivers, depending on the volume you’re dealing with, again, developer land versus pure B2B. I think volume is a driver when you start having inbox work, fully agree with that, then if you have a big signup flow and you work from developer signup going the path of converting or paying. If you start having like thousands of download, the inbox is going to explode.
If you’re pure B2B play and you’re trying to play with 1,500 customers, you can manage it. But I think there is that line, okay, my inbox is going to explode and I cannot follow on it.
The second one could be, hey, we hire somebody to do marketing full-time or we have a consultant half-time and we’re starting to be, I’m going to say sophisticated, such as we’re executing marketing campaigns and that could be another driver where you need tooling to do it.
Luke: For me the goalpost feels like when you hire your first salesperson.
Andrea: And that’s when we did it.
Luke: That’s when we did it at Clearbit as well. We’re kind of early-stage, you’re a bunch of people doing a little bit of sales and a little bit of marketing. But when we’ve hired our first real sales reps that was the time when we realized we should, time to bite the bullet.
Andrea: Right, and also two different people you want one place where it all lives.
Luke: I think the other thing is sales reps in general don’t like learning new tools, generally, and most of them are already using Salesforce. So their job is hard enough. You might as well not make it harder by forcing them to learn a different tool.
Peter: This is what we call “a moat,” in venture world. Yes, Salesforce stranglehold.
I want to shift topics a little bit to strategy and parsing all this data. You’ve told us that at some point you should invest in Segment and Segment’s going to help you get all the numbers in one place.
Let’s assume that at this stage you don’t have a BI team, you don’t have a full-time analyst who can write SQL queries. How do you actually extract actual information out of Segment or your data warehouse?
Luke: Clearbit is maybe a little unique in that most of the early team was fairly technical and SQL proficient. So for us, it was the right decision to put everything in Redshift via Segment. We used Mode on top of Redshift as our querying interface. There’s similar tools like Looker and Periscope that will let you run queries on top of Redshift, build charts and reporting and that sort of thing.
That was the right decision for us because we had the SQL proficiency on the team to make use of the actual data but I’m wondering if you might have some comments on how to extract insights for maybe a less technical team.
Nicolas: What I’ve seen, and even with what I’ve seen being successful, the big dream is to have that dashboard that’s going to refresh every day or every minute. You’re going to see your pipeline, whatever data and either you have great support which looks like you had, or I’m going to go back to a Google doc, and I’m going to explain why.
The big dream of adding all this data no matter what, if you go for your top five KPIs you’re going to have to pull those data from at least three systems no matter what. So the big dream is a unique dashboard and the data warehouse and the data mart and everything, is either you have the resource and everybody is super bullish about it, you build it, which is possible.
It’s not what’s going to happen, and I have a super strong point of view on this one by the way,
you’re going to run in circles trying to shoot for the perfect dashboard. And more or less you’re going to over-engineer it without ever getting the inside of it.
And even being at a company like HackerRank, hardcore developer, Twilio hardcore developer, when I started at Twilio I moved to a Google doc and we were like 20 people and we were trying to solve for the magical dashboard. But we moved to a Google doc initially. Why? Because we had to pulled from multiple systems, that was one.
The second one is there was no agreement on what we wanted to measure. So we started in a project and we had resources where I had one of the data analysts in Looker building a fantastic dashboard.
Spent like two weeks on it to pipe behind the scene working with the data engineers. Which most of the time normally if you’re small you don’t have but like working for two weeks and the execs looked at it presented at the team meeting and I’m like. beautiful dashboard, exactly what you guys asked.
First I had two minutes to present it at the end of the meeting and then they’re like, “Interesting, but we look at this once every three weeks.” And I’m like, “What? We’ve been working on this for a couple of weeks.” And they’re like, “Yeah, that’s not what we need.”
My learnings from that experience and I’ve seen it many companies is what do you want to measure and it will take some time to do two things. What to agree on and what you want to measure and I know after we’re going to talk about funnel so I’m not going to cover that topic.
The second one is to learn how to data, I think data are beautiful, you look at it but it takes you some time to appreciate the dynamic of the business and how your pipeline or your leads or your sign up are moving through the funnel.
What is pretty fascinating is it takes you two to three weeks to look at it. It doesn’t change much, actually, even when you go back in time except if you made a major change such as hired a salesperson, did a fantastic campaign or hire marketers.
But it takes that, I’m going to even go a month, like you said, it’s the dashboard a month to appreciate and understand and to have that level of conversation. So I’ll go Google doc, one month of learning and looking at the dynamic of the data, always look back.
Everybody says I’m going to look at current month I’m like no whatever go back in time, look at what happened between having nothing and something you’d better have something so I do not hesitate to try to go back 12 months even if the data are bad but just to get a sense and to get that sense about when we believe the data are better and what is trending.
It’s all about trend, knowing that at the end of the day what you want to do is you want to have your KPI, you want to have agreement and when you want to measure and after there’s a lots of discussion saying, okay, what is a benchmark? Let’s talk to the guy sitting next to me or another company about the benchmark. The benchmark initially is yourself.
It’s good to have a benchmark to see what’s going on, but what you want to do is you want to measure against yourself and improve whatever KPI you have.
And understand and being in control and when you are that level of in control and you measure against yourself and you have a sense about what’s going on then you move to the industry benchmark. That’s my pretty strong point of view, actually, in that part.
Peter: Luke, at Clearbit, reporting happens in SQL, sometimes using a tool like Mode. Nicolas, you’re a fan of using Google Sheets in the early days because you can iterate quickly. Andrea, what does reporting look like at LaunchDarkly?
Andrea: Again, we’re also using Google Sheets to bring lots of things together. But I know Nicolas does it as well. We have a lot of our campaigns live in Salesforce and a lot of them all drive to there. And by keeping them all in there we’re really able to see a lot of marketing effectiveness.
Peter: Again, we have some really early-stage folks in the audience. So could you just tell us what a campaign means in this context and how that lives in Salesforce?
Andrea: A campaign can be anything from just any event that you’ve done so if you have a way to connect people back to that event if that is a source of referral from a piece of content and you’re using UTM tracking, you can bring a campaign in like that. Anything that you’re able to say this is a piece that we did and we put out in the world and we hoped it to have an effect.
Peter: Great, thank you. I want to go back to the funnel. At this point we’ve got some preliminary tooling, we’ve got all our data in one place, kind of. And we have a preliminary understanding of what our funnel looks like.
We’re measuring signups, sign-up to product activation, product activation to paid, maybe even some downstream metrics. How do you decide what the first number to work on is? How do you prioritize work here?
Andrea: Recently our CEO Edith Harbaugh was speaking at SaaStr and she gave a really great talk. Look this up once the video is live, on “build versus buy.”
She said the biggest problem when you’re a new company is that no one knows who you are. Then the second biggest problem is that they don’t care. And really if you’re a teeny tiny company you’re not going to be able to affect a lot of different pieces within the funnel.
Look at top of funnel and know that maybe you will have to stop working on top of funnel campaigns for a couple weeks to build your onboarding process. Know that that means that you’re going to work on that and not your top of funnel.
If no one knows who you are you’re never going to be able to get them all the way through the funnel. I think you have to start there.
Luke: I think you underscored an important point. Make sure that you’ve answered the deeper fundamental questions about your business before you start worrying about building out a fancy funnel with lots of fancy reporting and lots of data.
Data is important but intuition is also incredibly important. You don’t develop that intuition about your business without talking to lots of customers and being really immersed in the world of your buyers.
Andrea: Take that idea of hypothesis-driven development and apply it to marketing. And think, “Great, I think that as a company that’s trying to create a category, we maybe need to do a lot of content,” or something like that. Have these hypothesis or have the hypothesis that this is your target audience.
But instead of rolling all of that out, find the first layer, the friends and family, to talk to. If you’re here at Heavybit you can talk to other member companies, which was really beneficial to us.
Then find the next small thing to do and keep getting that feedback on your messaging so you can be improving on that as you start improving your channels and finding the ones that are going to work for you.
Luke: Don’t underestimate the power of a well-written cold email, too. It works.
Andrea: If you’re able to articulate your value proposition, a cold email is a great way to get first awareness.
Luke: Right. Again, you don’t necessarily need to expect that the folks your cold emailing will buy from you but they will at the very least in many cases be willing to help or talk.
I think very early stage, maybe that’s your funnel, “How many cold emails am I sending or conversations am I having that are resulting from those?”
Peter: You two are telling me that top of funnel takes precedence over everything else? I do want to hear from you on this, Nicolas.
Nicolas: On this one you cannot solve for, the big dream is a lifetime value, actually. You want to solve for lifetime value. And no company solves for that. I mean, you have to be a 10,000 employee company to solve for lifetime value. But this is the dream and you want to build your campaign against lifetime value of a customer, so that’s not going to happen.
The point I’m trying to make here is, in a perfect world you have a funnel and you have multiple steps. What you want to solve for is saying okay, as a company what is success six months from now at the company level?
And then looking at that,This is what I do every time, it’s not like I change jobs often but every time I start a new job, which I did recently, six months ago. Or when I have the chance to talk to one of you guys, I’m like okay,
before you start explaining to me like all your marketing tactics, let’s pause for two seconds. What is success for your company six months from now?
Success can be many things. We want to have, to customer and production, and success has to be an objective with a number next to it. And based on that there is marketing playbook that you can decide and you can decide either to invest on top of funnel, mid funnel.
There’s so many things you can do, it’s overwhelming. But I think aligning to the company goal is number one. Yes, it will match to one of the stages of funnel.
I go back to developer land, but signups type of thing, your funnel is going to look like, I’m going to put the different stages but after, we’re going to show a couple. Website visit, signup, complete signup, complete signup is non-spam or people that are real human beings.
Because today there is a game when you are in the CS degree, the first piece of code you have to write is how you’re going to be able to create accounts on any website you find as much as you can. Which is a major challenge because you have all this data, if they decide to play with your website it’s going to be a little bit of an issue of having fake users.
Then there is a concept of product active, as I stated before, whatever you define as product active. It could be they went to a specific section of your product or they key something or they activated something push code… You have to define and then after there is that concept of put their credit card if it’s credit card business, and they’ve been spending over time.
Initially if you oversimplify that thing, you want to look at complete signup and product active by cohort. This will keep you busy if you have a high volume of inquiry, of signups. If you look by cohort by month it will give you some really good ideas. And I was doing that.
I did that exercise two times recently with somebody here and quickly and I think the feedback was, “I have no data.” And within a call of half an hour we’ve been able to see that he has improved the onboarding process.
Because he was highly focusing on it we’ve seen a higher conversion rate from complete signup to product active and some tactics and he was able to scope and understand what you should do next, aligning was his objective as a company.
If you look on the B2B side of the house, the frontal here is tender and there are benchmarks. You don’t have to use, but it’s enquiry, it’s anybody filling in a form. MQL is marketing qualified, whatever it means. You have to decide it could be based on persona, who they are, the company size, whatever you want and after there is a big step.
I’m going to skip SQL. SQL is sales qualified leads. It means people are either budget or a project to oversimplify the definition and then close. On this part of the funnel what you want to look at initially, because you’re not going to have scoring, you look at inquiries and SQL. You start there and to try to get a sense about what people are filling in and how many goes for that strong definition of SQL.
I’m recommending for people to be pretty strict on that definition because it’s a little bit of a gate telling you that, yes there is really something going on. The challenge sometimes was early startup and I’ve been coaching. I’m just sharing personal experience here. It’s mixing up pipeline with revenue and real pipeline with future revenue. So you just want to have that reality check thing here.
The guy is excited, meaning you’re more excited than he is which is a good thing. But it’s like, okay, is he really excited to the point of yes, there is a timeline or there is a budget? And then we can consider this as pipeline and after you can balance I fully agree on the cold email and everything that was said before.
But if you think hardcore funnel, this is what it is about. You don’t want to over-engineer it, you don’t want to overthink it.
I personally had the small startup less than seven people and the biggest mistake I’ve ever made is just looking at the spreadsheet all day long and getting excited and not talking to customers.
And I learned my lesson, it was a bad idea because you can do this all day long.
Luke: A funnel implies that there are customers moving through that funnel in the first place.
Andrea: Talk to whoever you can. Talk to anyone you can.
Luke: I think, too, we can also get carried away with SQLs and MQLs and it’s kind of foreign and sophisticated sounding.
But really, a marketing qualified lead could be someone signed up for your newsletter. A sales qualified lead could be someone fills out a form and it appears that they’re a good buyer for a product.
I think what you said, I mean, we don’t need to over-engineer or over complicate things initially.
Peter: You guys keep talking about SQLs, and Luke, you said that an SQL can either be a function of how a customer is using your product or an SQL can be determined by who the customer is. Can you tell us a little bit more about the qualification criteria you use for an SQL?
Luke: From the product users perspective, is this a customer that if you have a self-serve or a freemium model, are they using the product? If they signed up for trial they’re active in the trial, however you measure the success of that trial, they’ve cleared that threshold. With Facebook early on it was, have you added, was it three friends or five friends, that was the minimum threshold that they wanted to get everyone over.
But obviously on the other side of the equation, if the VP of Sales at a really hot company fills out a form saying, “I want to buy your product,” does it matter that they’ve had any product usage? No, they want to buy. They’re raising their hand saying, “I would like to buy your product.” They fit the target market you should go after that.
Peter: Some good strategies for lead scoring. Andrea, do you want to take this one?
Andrea: One is have marketing automation that helps you do it. There’s kind of two parts to lead scoring. There’s scoring by activity where every time they’re on the website or every time they fill out a form, they’re increasing the score.
And then there’s also the idea of scoring based on who they are whether or not there are part of your target market. But at LaunchDarkly we use marketing automation for that.
Nicolas: And after you can, he’s not going to say it, he’s too polite, you can use technologies for data automation. Clearbit does that.
And the point that is important in scoring is again you start simple. You have two dimensions, it’s behavior or demographic and based on your business and who you’re selling after we just went for that exercise at HackerRank.
I sat with the sales team, not for hours, but we sat together saying what is the perfect customer and who do we want to sell to title wise and company size and industry. Then we go with a v1 where not everybody agrees, it’s okay. It’s a v1, let’s chill, let’s implement it.
The first thing you want to do actually is not share, depending on the size of your sales team and so on, and depending on the volume, not share it initially and try to capture behind the scene and understand if yes/no is the score you put together, that translates into conversion.
If you don’t do that two-weeks test to look at it, what’s going to happen, let’s just call it self prophecy, is you’re going to tell your sales team, call them they’re better. Don’t call them there are worse and guess what happens. The one you told them to call are better because they didn’t call the other.
So you just want to validate your scoring program behind the scene. It doesn’t need to run for six months but couple of weeks.
Luke: I don’t think you need to get super sophisticated about lead scoring early on. In my opinion, you don’t need to come up with like a numerical scale between one and 100. Just think back to basics. Clearbit, our Salesforce integration drives a lot of revenue for us.
We can use Clearbit data to understand if someone is likely to have purchased Salesforce already so if someone signs up and they are already a Salesforce user, chances are they’re a good fit for our product and that’s someone who’s maybe a little more worth talking to than someone else who doesn’t fit that criteria.
Peter: Do you just have a binary yes/no lead score?
Luke: It’s kind of, are these valuable leads or not.
Peter: We have free, it’s pretty simple. It’s called sell simple, for sales people to understand. Three stars, calls them now. One star, call them when you’re done with the three stars. No star, don’t follow up.
Peter: How do you pull all these things together into one marketing campaign?
Luke: The way we do it, which is not necessarily the one true way, but we use a tool called Customer IO. The reason we like Customer IO is they support a fairly powerful templating language called Liquid..
What Liquid allows you to do is, since we have Clearbit data already and we use Segment, we pump usage data into Segment, Clearbit data into segment. That then flows into Customer IO. So we have a rich set of attributes for all of our customers in Customer IO. These are signups, people who’ve signed up for a newsletter, etc.
We have demographic details, we have product usage details and then from there we can build really crisp segments in Customer IO and even within those segments we can dynamically change messaging using the Liquid templating language.
Peter: We said that your two big choices for marketing automation tools are Salesforce’s Pardot or HubSpot. And we threw out Marketo as a third choice there.
Luke, you just said that using a different tool, CustomerIO, when did that fit into your roadmap?
Luke: We started using CustomerIO pretty early on. It was definitely in use before I joined and I was employee number seven.
Peter: Andrea and Nicolas, how do you guys feed product usage data into your email automation system?
Nicolas: That’s the dream, because the level of insight you get is pretty fantastic. I’m going to go, at LinkedIn, I never had it, into CRM, just to be clear. At Twilio, yeah not better. And here at HackerRank, I don’t have it either, feeding into CRM.
So the way we had to work is to build a data mart where we would feed the other way around. Data mart would be at the core, you would have product data flowing in, you will have CRM data and marketing automation data in one place.
It was a piece of art. You really had all the data to do segmentation and targeting and it took us some time, it took us six months as a project. Super-scary, everything I said I’m going to go back to a simpler version.
Took us six months as a project to be at that level of sophistication where we spend the time and the energy to get all the data. There is a work-around, this is what I’ve done at Hacker Inc for now. It’s not great but that’s as good as it gets.
I’m getting an export through FTP for now off Redshift data from the product with email address, signup date, active date and some other field that I manually matched through Markerto, into Marketo that pushes it into Salesforce. And based on this the marketing campaigns are built behind the scene.
I’m sure CustomerIO is better, we don’t have Segment, but that’s what we did. By the way, to get started those type of manual imports work nicely but it doesn’t scale, just to be clear. Because you cannot on the fly.
It’s like you have a picture at one point. I know it works because I’m a big data guy and I measured behind the scene of, hey, if somebody is engaged on product, how is your conversion rate? Yeah, it’s a no-brainer. It works. But at scale, I haven’t solved for it in a simple way yet.
Luke: It’s the Holy Grail and it’s hard to get to. I think until the perfect marketing automation tool exists…
Nicolas: Which doesn’t exist.
Luke: Which it does not, never underestimate the power of a CSV export in Google Sheets. I think a lot of segmentation happens by hand. Exporting data, munging it, cleaning it and building those lists. And then some sort of mail merge tool. Nothing wrong with humble beginnings.
Peter: Andrea, I see you nodding. Are you running campaigns based on how customers are using LaunchDarkly?
Andrea: On some level. We do use Segment to bring certain pieces from production app into Pardot and Salesforce. So we have some information but there’s a lot of levels of sophistication we have left to go.
Luke: I think there’s no tool that out-of-the-box is going to get product data into your marketing automation tool. No new tool that I’m aware of.
Nicolas: Listen to what we did.
Luke: Even our Segment integration, we have a series of scripts behind the scenes that are pulling the data from our product and sending into Segment. So there’s definitely customer development work that we’ve done on that end.
Nicolas: And it’s worth it. The thing I really want to have as a message is it get quickly marketing and I was discussing it with somebody else just before on another meeting. Is overwhelming, you can do so many things like this.
You can work on a project like this for months and you will see progress really having like this, “Hey let’s pause, let’s see what we want to solve for. Let’s see how much time it’s going to take and being sure that we do the right thing is really important.”
I do not believe, I was sharing this on the prep call that there any technology gap or as you’re right actually there’s one, there’s no that perfect tool that pull but you can buy so many technologies.
The issue and the major challenge is agreement in the goal and process. That’s the issue.
Once you’re clear of what you want to solve for and there is agreement on whatever it is, a funnel, a thing like the measure of success, then it will get easier. You just have to focus, which is always tough, and start a plan because there are so many things you want to touch.
But it’s really, and I love technology, I was running marketing operations for all of LinkedIn worldwide. I had B2B business, online or credit card and consumer. So I would have I have a team of 100 people and every day I would have somebody saying, “We want to buy but then I have time I’ll do it on the side.” I’m like, yeah, we’re not going to do that.
But the point is even whatever scale you are it never stopped, it never stopped. And I don’t think there’s a perfect approach to it but making those right decisions is really really important, I think.
Peter: Tools to monitor conversations outside of your existing channels. How do you stay on top of Twitter, Reddit?
Nicolas: Nothing good on my end.
Andrea: Hootsuite would be one just for Twitter.
Luke: Yeah, we use Zendesk as a social monitoring tool for the Twitter front, on the support side also has some social monitoring tools.
Nicolas: There is a cool tool, something we’re exploring. I don’t even remember the name of the vendor but no there are bots that can replace SDR. I love SDR, I appreciate them but the point is they don’t follow up on everything, sales development rep. And now there’s, it’s Conversica.
So there is a new technology we’re going to play with where more or less, you just throw your leads there and the bot that you can name as you want, Bob, Eric, Olivia, Alexandra, will start engaging with the prospect and is able to understand like what’s going on, ask questions, and it means like nobody gets untouched.
All the big discussions we had on scoring and everything you can kill part of it. You activate that bot and then based on a threshold you can reroute this to the salesperson. It’s something new, I think it cost a bit of money so that’s where I’m getting stuck. But the point is, people are even creating a LinkedIn profile. I would not name the company who did that but more or less a person has a LinkedIn profile, there’s a picture and everything, her name is, what, Alexandra?
You can go pretty deep now for that automation which is pretty cool because one of the big challenges, one is do you have the resources to catch up with the lead flow and then the qualifications, the MQLs, will solve for all of it. Because if you have a certain level of engagement, you can say, sales, follow up on it. Pretty excited by that.
Peter: What’s the distinction between a sales conversation and a marketing conversation with a customer? Andrea, you’re smiling.
Andrea: Yeah, I’m just trying to think about what that would be. Certainly feedback from sales is just as important as well.
For me a marketing conversation is, what was your journey to finding us? What was it that brought you in the door and what convinced you that this was going to be the right solution for you?
Really just hearing that story and then also kind of trying to put together some of these hypotheses that you have about for us, it’s have you feature flagged before, therefore if you’ve already done feature flagging, therefore you’re going to be more likely to buy us.
Things like that and just collecting those stories over time and starting to create those relationships that you’ll also need for case studies and things like that. So you always have somebody to say “I’m curious, what do you think about this?” Before you release it to the whole world.
Luke: Yeah initially I’m not sure if there’s much of a distinction. I think you’re just talking to customers initially, right? You are probably the sales and marketing team at your startup. So I think further down the road I think there’s of course more distinctions there.
Nicolas: If I jump on a sales call, most of the time I would not say a word. I’ll be a ghost and what I am interested in is how is the discussion going and how is a salesperson doing the pitch. I’ll do five or 10 of those in one week. I try to stick with the same sales rep, meaning it depends on the size.
Why? Because otherwise everybody has their style and everything but just more or less I go to the VP of Sales and I’m like, okay, where is the two best ones, and I’m going to go there and hear the output of those discussion for me as I’m going to be able to work on the pitch deck and train the sales team based on the feedback and how I believe they should do discovery, a pitch and so on, after it’s up to them to use it. So that’s one part.
The second one I think a one-on-one with a customer would be more from a product standpoint. Think product marketing a product and it’s discussing and more going deeper in the use case and what they are solving for.
Even if they decide even on closed/loss deal you can do it on close/loss because what like existing customer, customer you lost against and here is we need to understand the connection between the use case, the market, and the product you’re building and again to try to do many of those.
No matter what, I have never been disappointed on anytime I met with a customer. You learn something every single time.
Peter: Been a big theme during this panel, “talk to your customers.” On that note I’d like to thank the three of you. This has been really educational. Please stick around, some of us will be hanging around. If you’ve got further questions, come find us.
Interested in joining Heavybit? Our program is the only one of its kind to focus solely on taking developer products to market. Need help with developer traction, product market fit, and customer development? Apply today and start learning from world-class experts.