B2B data + AI hype: Q&A with Seismic's VP Data & Analytics

Many of us in the digital marketing / martech industry know all too well the hype around AI. We also know the challenges of dealing with data in the B2B space.

For many of us, we see a lot of “AI-powered” tools on the market. How can we distill the noise to get our problems solved? What kind of data analysis should we be doing ourselves, versus using a BI tool for?

For those who are new to data analysis, where can we go to learn?

We sat down with Al Bsharah, VP Data & Analytics at Seismic, to pick his brain on exactly those questions. 

Side note, Seismic became a unicorn in late December. A megaround of $100 million Series E funding pushed them over this milestone. They are now valued at $1.05 billion.

As Al said, “it’s been a lot of fun.”

ClickZ: Tell us a bit about your background?

Al Bsharah: I started as an electrical engineer in the Detroit auto industry, and made a career out of that for a couple years. Then I moved to San Diego and got involved in startups. I was the second employee at the first company I joined, which was acquired. Since then, I’ve had a couple other companies.

The most recent one, Email Copilot, we sold partly to ReturnPath and partly to Seismic.

CZ: And now your current role is VP of Data & Analytics at Seismic. What’s your day to day look like?

AB: So my first year at Seismic was very tactical. Building technologies in the right way to last us for the foreseeable long term future.

Then the next year we were very focused on strategy. Where are we going, what are we building, what kind of technology should we be leveraging inside our platform and how do we get there? How do we build a good search on our website, how do we get relevant content?

This coming year will be very much focused on product strategy. Similar to last year but more focused on the product itself. How do we think about our client’s workflows and the problems they’re trying to solve? And how can we build intelligence across those?

CZ: Seismic does sales enablement and marketing. That’s a lot of data to cover.

AB: Yes, it is. The area is very wide and deep. It covers a number of personas, from sales enablers to marketers to sellers to buyers. Thinking about the multiple workflows people have to go through just to do their jobs.

We try to provide intelligence around a specific workflow.

How to decide what content to use for what? At what stage of the funnel? Through what medium? That content is being used for different personas.

We try to deliver a heat map of sorts showing that content throughout its entire lifecycle and where it works.

That’s how I’m kind of thinking about data, especially in the B2B world.

CZ: When working with B2B vs B2C data, what specific challenges do marketers face?

AB: To start with, there’s just less data. But there are pros and cons. The data has a higher value per item. It’s more impactful if you get it right. But it’s also more impactful if you get it wrong. You have to be very careful there.

B2B data also has sparcity issues. If you want to do testing at any level, if you try to slice that data down at any level and it gets thin, that can be challenging to build insights from. At least insights that are actually reliable, something you can believe in.

In B2C, you might have millions and millions of users. This is just a guess, but Spotify might have 100x or 1000x the daily active users that Salesforce does. 

You have to invest really heavily in gathering and cleaning the data, so it’s a much higher quality. Versus in the B2C world where there’s just so much more data to use. 

So there’s a lot of effort in trying to reduce the amount of labelled training data that’s needed. You usually have models that are training data, and putting labels on it so that a machine can learn from it. With B2B, we need to put effort toward reducing that amount that’s needed.

Ultimately, one of the biggest challenges in all of this is maintaining trust. Because of the higher value and lower frequency of sales and interactions, there’s less room for error.

If we’re making big predictions on how do we engage with this certain buyer and we do something wrong, that could be very painful.

So for us at Seismic, this is a huge challenge to be able to take data and turn it into something that is actionable and usable for the B2B space. It’s fun and it’s exciting. But it’s very hard. 

It’s hard because very few people in this world have the skill set to make that translation. It’s also hard because every marketer has different needs. You can’t solve every problem for everyone.

How does a marketer get a little more technically adept so that they can answer some of their own questions? Because they may not be getting the answers they need. Maybe it’s hard to translate what they have into an answer.

CZ: Or maybe they don’t even know what questions to ask

AB: Oh yeah, absolutely. And a lot of times the people delivering those solutions don’t know the questions either — and they’re trying to figure out how to solve them.

CZ: What can marketers do to start learning about their own data?

AB: So for marketers, try to get access to whatever data is provided to you in a raw format.

Start by just throwing it in a spreadsheet. Excel, google sheets, whatever. Learn about pivot tables, play with some formats. Get comfortable with some tools that tell you about data. Build a chart or a graph that might be beneficial to you in your daily life or work.

Maybe play around with that for a while. Then you graduate to a BI tool like Power BI, or Tableau, something like that. The point is that you’re able to glean some basic insights out of your data.

You may already be using a tool that tells you about your data.

But it’s your data. It’s an important skill to have.

Get into it, get uncomfortable. You may not know how to use excel or how to do a pivot table. You may not even know how to get your data.

Figure those things out. Start plugging things in. Start trying to answer some questions that have been bugging you for the past 12 months. Naturally, you’re going to stumble across some things. You’re going to have some ah-ha moments.

In a few months, you’re going to be not just a better informed person for your job, you’re going to be a more valuable resource for your company.

CZ: That’s a really important point about owning your data. So say someone is at the intro level to the world of data and analysis. Where should they go?

AB: I don’t have a definitive answer, but there are a number of different paths you can take.

Some people say, “show me the ropes.” To them, I would say go online to places like Udemy or other online courses, and find a course you want to learn. There are great courses on Excel, on Power BI.

They’re relatively inexpensive, or sometimes free. If it’s related to your job, I bet your company would be willing to expense the low cost of some of these courses so you can be educated by an expert in the field.

Other folks are more DIY / figure it out yourself. So for them, maybe they should just play around with the tools. Maybe they’ll hit some roadblocks. So they’ll google what they’re trying to do, find the answers, and move on to the next thing.

We live in a world where if you can’t figure out how to educate yourself in a number of different paths, you’re probably not trying hard enough. There’s so much information available.

CZ: For digital marketers in general, is this a skill everyone will need to pick up on versus relying on tool to do it for you?

AB: You have to at least understand what kind of data is out there. What data is available to you? We’re moving in a direction of “data is everywhere.” The B2B space especially is now getting more aware of the fact that, “holy cow, there’s a lot of data here.”

You have to have at least an awareness. You have to be able to think about those things.

If you’re evaluating a new tool or technology, you need to keep up with the types of things the industry is doing, or the data that’s available, or how certain technologies can learn certain data to do things.

For example, if a certain technology is able to understand a certain page that a prospect is looking at, which content, and for how long — if you don’t even know that’s possible, then it’s hard for you to understand what can be done.

That information is valuable because if I know what page they’re looking at and what’s on that page, then I know what they care about most, so I know what to talk about with them next.

You can’t make that connection if you aren’t paying attention to the data you have.

CZ: Say you’re in the market for an AI-powered martech tool. What are different red flags or must-haves?

AB: There are a lot of red flags. I think the AI craze has been out of control for a while. And it’s set the bar ridiculously high for what kind of tech even exists right now.

It’s set some misconceptions around not what’s possible, but what’s reliable and what’s accurate and what works well.

With B2B, again if you get it wrong, that’s a big problem.

My thinking right now is that there are a lot of cool technologies doing a lot of cool things, but there are also a lot of technologies doing a lot of mundane things.

I forget the exact stat, but sales reps spend maybe 35-40% of their time actually selling. The rest of that time is doing administrative stuff, trying to find content, updating the CRM.

I think what people are thinking about is, “Oooh look at this tool that helps my sellers in that 35% of selling.” And that’s great, if you found a tool that does that and it’s reliable. But oftentimes, it’s risky.

The far less risky thing to do is find a tool that deals with the other 65% of stuff that sellers don’t want to be doing anyway.

Leverage the tools that automatically update the CRM, for example. Or help set up meetings, things like that. So that your sellers can spend 45% of 55% of their time actually selling. That’s what they’re good at, that’s what they do best.

CZ: If you were in marketing or sales, what kind of tools would you rely on for your work?

AB: There are a lot of great automation tools out there. Automate the mundane.

These are hugely valuable, they free people’s time up to do what they got in this field to do in the first place. Writing better content, or selling to a prospect. Give them the time to do that kind of stuff.

You want to cherry pick some other things too. You don’t want to be completely conservative, you want to have your head in the game. Some stuff is really advanced. Doing analysis of meetings and sentiment, are people speaking the right way when they’re selling. A lot of cool tech that helps sales managers better train their teams, based on what’s actually happening in their calls. There are really cool predictive technologies on what should you do next, and what deals should you pay the most attention to.

Find the happy combination of what you’re comfortable with between mundane, less risky stuff that allows you to do more of what you’re good at, and the cutting edge things that may or may not be 100% accurate. You have to be okay with the risk, and keep a close eye on that.

CZ: How can we sift through some of the noise / hype around AI?

AB: You shouldn’t be focusing on “oh, do you have AI?” I think the industry is kind of shifting in this direction right now. We’re a little past the peak of the craziness.

Now we’re heading more in the direction of “okay, I’m not sure I care if you have AI or a better UI. What I really want is for you to solve my big hairy problems. If you can solve my big hairy problem, I don’t care how you do it. Just help me do that.”

If we all just focused on that instead of some fancy predictive algorithm, I think everybody would be in a better place. That’s what matters. If you’re not solving one or many of the problems that users have, then the rest doesn’t matter. Just make that easier or more efficient for them.

If we’re not solving problems, we’re wasting our time. That’s why we’re here. That’s why our company’s here, that’s why every other company in the world is here. It’s why we all do what we do, we’re trying to solve problems. That’s got to be the focus.