Want to drive more predictable growth in 2025? Start by looking backwards.
That’s how Megan Bowen (CEO of Refine Labs) and her team start every engagement with a client. If you’ve been on LinkedIn, you’re probably familiar with Refine Labs. They primarily work with B2B SaaS companies ($20M to $1B in revenue) to optimize paid search, improve paid social, and drive more qualified conversions.
Last week, Megan came on the podcast to share how they leverage historical data to help their clients get more customers. When they start working with a new client, the first thing they do is conduct what they call a “revenue performance assessment.” Here, they analyze historical data to find growth opportunities and create more accurate forecasts.
Read on to learn how it works and how you can do it at your company!
Or listen to it on Spotify or Apple
First, gather 12-24 months of historical data. Megan says this is the sweet spot – if you go back further, you might run into outdated data that’s no longer relevant (like the COVID boom of 2020-2021). If you only have 2-3 quarters of data, that’s better than nothing, but you might miss important seasonal patterns.
Focus on two main sources:
You’re specifically looking at these key data points:
Once you’ve gathered each of these data points, it’s time to start analyzing them.
Here, you’re viewing all closed/won by pipeline source to understand how different pipeline sources convert differently. The goal is to understand which of your efforts are moving the needle on pipeline, and which aren’t.
When it comes to attribution, Megan says there’s no perfect model. “Marketing isn’t intended to be measured perfectly. And that’s not the goal here. The goal is to identify trends and come to hypotheses of what we think the data is telling us. It’s never going to be perfect. And when we think about pipeline sources, we’re looking at sort of the tipping point of what influenced that lead or contact prior to the conversion. But we all know that there’s a million things that that person experienced or interacted with before they came a lead.”
Megan recommends blending “software-based attribution” and “self-reported attribution” to gather signals of what’s making a meaningful impact on customers who end up buying from you.
Software-based attribution is great for measuring “demand capture,” what channels you’re actually using to capture the existing demand for your business. Examples here would be paid search, someone coming direct to your website, or people searching you on Google.
Self-reported attribution (asking, “How did you hear about us?” in your main conversion point) helps you understand how the demand got created. For example, software might tell you someone came direct or searched for you on Google. But what drove them to type your domain in the browser, or your name into Google? Here, you’ll get a sense of some of the things that helped create that demand: your CEO’s LinkedIn posts, team podcast appearances, word of mouth, or a specific marketing community.
With this data, you can do a “split the funnel analysis.” Here, you look at how different pipeline sources convert through your funnel. This helps identify:
For example, they may find a company running a direct response LinkedIn ad that’s generated 300 leads but only ended up converting into 2-3 opportunities.
The next thing you can do with your historical data is analyze your wins and losses to understand trends, validate your ICP and targeting, and see which segments are especially valuable for you.
For example, you might be focused on bringing in more leads from enterprise prospects, but 70% of your closed revenue comes from mid-market. Or you might discover you’re crushing it in manufacturing while struggling in other verticals.
As you go through this audit, you’ll likely find that you’re missing some key data points. Maybe you struggled to analyze closed/won deals by pipeline source because your attribution isn’t set up quite right. Maybe you only have 9 months of historical data. Or maybe your CRM has significant gaps.
If so, you’re not alone.
“Everyone has some gaps in their data,” Megan says, “Nobody has a perfect CRM. Part of this process also, little tangent side note is, as we go through this, we also try to identify where we see data gaps exist. And that’s actually a benefit of this exercise because it’s an opportunity to not only do the analysis, but identify where you can be tracking things in a better way so that going forward, you can improve just the data accuracy that you have.”
So as you go through the process and come across a “data gap,” make a note to go back and fix it later. Some of the most common gaps Refined Labs sees are:
Again, no company has it perfect. So don’t let perfection stop you from starting this process in some form now, and improving it incrementally over time.
Once you’ve analyzed 12-24 months of CRM & automation data, analyzed inbound conversion flow, conducted a “split the funnel” analysis, and analyzed closed/won data, you can use these insights to drive future growth.
First, start with a baseline. You can forecast future growth by taking historical performance and projecting it linearly, accounting for seasonality. And thanks to your “split the funnel” analysis, you better understand which channels or programs are driving most of that pipeline.
Next, identify growth levers by looking at conversion rates across the funnel. Specifically:
For example, you might find, “Wow, we really need to improve our conversion rate from form complete to opportunity created. Something is wrong there. If we just convert 10% more of our inbounds to opportunities, we can hit 20% more revenue next year.”
And now that you understand what channels or programs are driving the pipeline, you can better model different scenarios:
What if we spend more on advertising?
What if we improve specific conversion rates?
What would the impact be on results?
The key here is to stay focused. “Don’t try to boil the ocean. Don’t try to have like 10 things that’s gonna help. Just zero in on the 2-3 things that are really gonna move the needle and focus on those for three to four quarters.”
Common levers Refine Labs find include:
Databox makes it easy for everyone on your team to analyze their historical performance, or build dashboards to do your own “split the funnel” analysis. It also helps you answer some of your most important questions, with forecasts, benchmarks, and AI performance summaries.
Watch our tutorial to learn how to:
Get practical strategies that drive consistent growth