At Parabola, we’re building out the team, processes and tech stack to support our next phase of growth.

We recently revamped our messaging. We adjusted pricing and packaging. We introduced a new sales motion. And we dramatically narrowed down our ideal customer profile (ICP) in a data-driven way.

Throughout the process, almost all of the individual projects felt great. Killer webinars, awesome case studies, an active operators community, dinners, thought leadership, and (much) more.

That said, when you zoomed out from an individual project, it was clear that these efforts were not driving meaningful ICP pipeline – despite the fact that we knew exactly who we were selling to and what we needed to say.

We were virtually running “random acts of marketing” for months – hoping that these acts were hitting our ICP instead of focusing efforts on a small group of accounts and seeing what actually influenced them.

With pipeline growth clearly being our biggest bottleneck, our CEO gave the team a mandate: build a system that could create consistent and repeatable ICP pipeline within 30 days.

Here’s exactly what we did to bring that system to life.

Acknowledging the problem

One great thing about selling a data automation tool is that we have a GTM team full of automation experts. As a result, we have Slack alerts set up to notify us about almost every action you can imagine – a target account visited our pricing page, a prospect opened an email 5 times, a trialist ran into a product limit, etc.

The problem was that this data wasn’t flowing into Salesforce in a way that connected these activities and signals to one another. And without that unified data, we didn’t know where in the buying journey our target accounts were getting stuck.

The accounts we were going after weren’t totally random – we did have some semblance of an account list. We knew these accounts were a good fit and we knew that they had engaged with Parabola in some way, but we didn't know the quality of that engagement or which follow-on marketing strategies would best resonate with them.

If you were savvy with our stack and had an open afternoon, you could potentially back your way into a list of engaged accounts. But it would’ve taken you days to figure out what we were considering our TAM, how many accounts were at each stage of our ‘funnel,’ and what needed to happen to push a meaningful number of accounts to the next stage.

All the while, we were telling ourselves that we were running “an account-based motion” despite everyone having a different understanding of reality. We really had to come to terms with this dissonance when an investor (who may or may not be named Kyle Poyar) asked us, “Who are you trying to get in front of and where are they in your funnel?” and our answer was less-than-stellar.

After a couple long days flowcharting the process, every team quickly developed a shared base truth for account stages (in our world: Identified, Aware, Interested, and Evaluating) and definitions. Next, it was time to build out the supporting infrastructure.

Building reality

At this point, we knew which data points we wanted to take into consideration when determining account stage. For instance, we wanted an account to move from Identified to Aware if they visited our website, joined our ecommerce/logistics operators community, or attended an event (happy hour, webinar, etc.)

If you’ve ever spent meaningful time building out complex workflows in Salesforce 1) I’m sorry and 2) you can appreciate the complexity of this build.

Since we don't yet have RevOps in house, we initially tried outsourcing the expertise. As great as they were, the experience ended up being quite painful – to them, this work was little more than a set of tasks, and communicating the ‘why’ behind every decision was crucial but near impossible. As a result, we lost a ton of precious time QA’ing the work and going back and forth with the consultants (remember – 30 day deadline).

While we got to where we needed to be, we had a serious learning – we (ie. the team who truly understood the details) needed to be able to control the foundational logic. The motion will surely evolve over time, but we didn’t want to build an even greater dependency on others this early.

Controlling ABX stages

We drank our own Kool-Aid a bit with this one, choosing to build the ABX stage workflows with help from a process automation tool called Parabola. If you’re unfamiliar, the TL;DR is that Parabola lets you build logic-rich data automations without writing code.

To give you a sense for other infrastructure we have running on Parabola – we have one automation pulling in all email and call data from Outreach and storing that data in a Parabola-hosted database, and another automation pulling from that database and piping email open data into Slack and Salesforce.

Building this in Parabola ultimately allowed us to scale the logic and make it so that anyone on the team could open up the automation and read through the process documentation.

For instance, if someone wanted to see why an account moved from Identified to Aware, they could read a card, click into the Filter logic, and see exactly how and why an account was updated:

We ended up having two workflows in place for simplicity: one built to advance accounts down funnel and another built to regress accounts if a certain amount of time has passed without advancing.

Now, when someone asks, “How many days can an account be considered Interested until they move back to Aware?”, they can easily find their answer by clicking into the card.

After finishing and quadruple-checking the automations, we knew there would inevitably be mistakes, new logic to add, and aspects of the process that we overlooked.

The first major oversight popped up days after setting the automation live when we realized we had nothing in place to move an account from Evaluating to Customer 😅

Within 20 minutes of identifying the process gap, we were able to update it – adding logic to pull in recent Closed Won opportunities, checking if any of our Evaluating accounts were associated with a Closed Won opp, and updating their stage if so.

The update didn’t require us filing a ticket with a Salesforce admin or consulting firm – rather, as soon as someone on the Marketing team identified the gap, they were able to jump right in and make the update themselves after getting the ‘okay’ from the team.

It’s hard not to get excited when problems are being solved nearly as quickly as they’re being identified. Not to mention when your team truly owns processes end-to-end, it means you’ve actually thought through all of the details and implemented things exactly as they were dreamt up.

The point is not to use Parabola to build your ABX logic, but rather to state the value we’ve seen from building your source of truth in a place that’s highly visible, well documented, and easy to iterate on when you inevitably up-level your ABX approach.

That said… if you do want to test Parabola for ABX automation/reporting, here’s a free starter automation you can use in Parabola.

Aligning GTM efforts

With stage definitions, accurate account- and contact-level data, and our stage-change automations in place, we finally had clarity. Anyone on the team can now tell you:

👉How many accounts are at each stage

👉Why every account is at a certain stage

👉What tactics we’re running on an account with the goal of advancing them to the next stage

One culmination of this work came when we showed Scott Maxwell (OpenView’s founder who pushed us towards this motion along with Kyle) our final flow charts and data breakdown. Candidly, we were a bit embarrassed about the grand reveal because of how few accounts were in the Interested bucket at the time.

His response was along the lines of, "This brings a tear to my eyes. You now have a true ground truth."

He was fired up despite the Interested numbers because of how quickly the strategy was stood up, and how it meant we could finally operate from shared truth and start addressing gaps in our marketing funnel.

ABX data to data-driven experimentation

With new visibility into accounts across stages, one massive theme quickly became clear: we had a ton of Aware accounts (largely based on email opens and website visits), but very few Interested accounts.

As we thought critically about the why, we realized a few things:

We needed to broaden our entrance criteria for the Interested stage

Since the criteria was initially limited to events and trial signups, we weren’t properly moving accounts that had multiple visits to our pricing page into Interested (for example)

We needed to start running and learning from tactics specifically designed to move accounts from Aware to Interested

The first coordinated campaign we designed focused on this specific stage progression was about how teams like Flexport and Brooklinen are using our AI features to extract messy data from PDFs, emails, and spreadsheets across their supply chains.

After aligning on the theme, pain points, messaging sets, and use case examples, the team began spinning up tactics with explicit tests to help us better understand how to move accounts down-funnel. High-level tactics and tests included:

-Scaled email: Do we get more positive replies by offering to share a 2-min video or meeting them in person? What about if we include a link to our video in the email vs. asking them to reply if interested?

-LinkedIn automation: Are positive replies impacted by having the email sender follow up with an automated connection request, message, and voice note? Does that change if the automated LinkedIn touches come from another leader on the team (ie. my teammate mentioned he reached out)?

-Organic social (thought leadership): Does a text- or video-based post perform better? Within video, is a product-focused video or ‘talking-head-style’ video more effective at driving organic impressions and post engagement?

-Paid social (LinkedIn): Is a post including a screenshot of a customer Slack message more effective at driving engagement than a post boosting our most notable reference customers?

-Nearbounding: Which CTA is most effective at getting positive replies? What degree of personalization is necessary? How big of an impact does omni-channel outreach make on our meeting booked rate?

-Community engagement: Are our community members receptive to engaging with our CEO’s LinkedIn content when we bring the post to a Slack thread?

Within all of these, the team is clear on experiment details and metrics we’ll use to quantify success.

As we approach the end of this campaign later in September, we’ll be digging into the data and walking away with learnings to inform our next campaign. Next up – seeing this foundational work actually translate to repeatable revenue.

From random to unified GTM in 30 days

Adam Reisfield

More Techbible