The future of SaaS lead scoring is here - and it’s just getting started
We need a new way to lead score. For a long time, I’ve thought - and I think most go-to-market teams operating in the SaaS world would agree with me - that the traditional framework for SaaS lead scoring doesn’t do anything at all to help businesses with acquisition. The points are allocated at random. There’s little to no correlation between someone downloading an ebook (50 points) or attending a webinar (75 points) and intent to buy. Content consumption is about building brand awareness and affinity. Reading a blog is not a request for a demo. Sales teams know this, which is why they don’t trust the list of leads with high scores passed to them by the marketing team. And - although they’re probably too scared to say it out loud - most marketers don’t really trust this method of lead scoring either. But they keep doing it because there’s never been a better alternative.
That’s changed with the advent of SaaS product-led growth. Now, SaaS businesses are offering free trials (Ortto included) or so-called ‘freemium’ software (think Canva, Slack, Invision) to generate their leads. The traditional marketing funnel has been flipped upside down, making the old-school approach to lead scoring not just ineffective, but completely redundant. These days, SaaS businesses are letting their products do most of the selling for them. They’re taking a new approach to lead scoring too: they’re still looking at demographic and firmographic data to help them identify prospects who fit their ideal customer profile, but they’re combining it with data from their actual product usage - rather than visiting your website a few times - and leveraging machine learning and AI to identify the prospects who have the highest likelihood of converting.
While this approach to SaaS lead scoring is leaps and bounds ahead of where we were a few years ago (and, let’s face it, where many businesses remain today), I don’t think it’s finished evolving yet. In the future (which might not be too far away, given the lightning-fast progress of AI of late), AI neural nets will not only predict which prospects are most likely to close, they’ll tell us who our ideal customers are, too, finding patterns in data that humans could never see. And you know what? I can’t wait.
Why traditional SaaS lead scoring doesn’t work anymore (and perhaps never has)
Traditional lead scoring is fundamentally flawed. But it is also foundational to the way many B2B go-to-market teams generate leads.
“For more than 20 years, the MQL has been the focus of most (if not all) B2B revenue processes,” writes Terry Flaherty, VP and Principal Analyst at Forrester. “Leads are scored based on a combination of profile (explicit) and engagement (implicit) factors, and those leads that reach the scoring threshold magically become MQLs that are sent to either the RDR/BDR for qualification or are sent directly to sales. It’s been this way since the late 1990s, and it’s fraught with problems.” At the heart of the issue, is the fact that this scoring takes place before a “lead” has had any sales-driven engagement with a brand at all.
Most marketing and sales teams know that this approach to lead scoring is unreliable at best and irreparably broken at worst. After all, they can see it in their conversion rates. Research from SiriusDecisions (now Forrester) found that 98% of marketing-qualified leads (MQLs) never result in closed business, while a survey by OpenRise found that only 35% of salespeople showed full confidence in their ability to accurately score leads. Even fewer (32%) said they were very confident in their organization’s account scoring strategies. Despite this, many go-to-market teams persist with lead scoring they know doesn’t work simply because they lack a better alternative.
While working at Zendesk, Guy Marion, then the platform’s Head of Sales, decided to investigate the efficacy of Zendesk’s own lead scoring mechanism. His team ran a quarter-long experiment where they selected 400 “ready for sales” leads and 400 non-qualified leads at random. The goal of the experiment was to understand how much better the scored leads performed when engaged by sales. “Despite our best efforts, we found no statistical difference in our ability to connect with, re-engage, or win the “ready for sales” leads compared to randomized, non-scored leads,” writes Marion.
Content consumption doesn't correlate with intent
Wes Bush, the founder of ProductLed, writes about what he sees as the flaws of a traditional lead-scoring approach in his book Product-Led Growth. “It encourages marketers to gate content to hit their MQL goals. It focuses on content consumption as a leading indicator of intent. The entire process rewards creating friction in the buying process. As a result, there is often a disconnect between marketing and sales. Should we really be surprised? Does downloading a whitepaper mean you’re ready to buy? Absolutely not.”
Phil Vallender, Director of Blend, a B2B marketing agency, agrees that there is little correlation between downloading a whitepaper and a desire to book a demo - at least, not in the immediate term. He notes that the theory behind this approach - the idea that a prospect who fits your ideal customer profile and visits your content is therefore ready to buy,
but they just don’t know it yet - is ultimately misguided. Content, after all, is a long game, designed to build brand authority and affinity over time, rather than being a short-term play, and in fact, operates in the opposite way to the entire notion that traditional lead scoring is built upon. A content marketing strategy is designed to attract and nurture ideal customers over time who aren’t ready to buy so that when they are looking for the solution your business offers, your brand stands out from the competition.
“Traditionally, the idea is that someone fills in a form and downloads a piece of content, then another piece, they visit the site so many times, and so on, and you stack them up until they reach a score that’s on par with someone who has actually put their hand up and made a demo request,” says Vallender. “But that’s just not the case.”
Implicit bias in traditional lead scoring frameworks
A further challenge to the traditional lead scoring model is determining the right score values to assign to leads based on their behavior or demographics in the first place. “This is really an attribution exercise that is best handled by statisticians and software, but typically falls to marketers who have to act on gut feel and assumptions,” writes Marion. “Without the support of a statistician or personal experience in performing and interpreting multivariate regression analyses, marketers can’t possibly know how important each of the above signals is in predicting whether a buyer wants to engage sales and/or will buy. Even if they did know exactly which signals drive conversion, they can’t weigh up the relative importance of each.”
Even with the benefit of that statistical or development expertise, “a human-built scoring model is subject to the biased beliefs of its developers,” notes David Campbell, Vice President of Product Marketing at Applied Systems. “The people building the model select the attributes and engagement actions of a lead score, and the relative weight of these attributes and actions.” Intent indicators must be measured for impact to conversion, continues the report. “However, while many intent indicators lead to conversion on paper, they may have very little correlation in practice.”
Mike Sharkey, CEO and co-founder of Ortto, adds that “in addition to just how impractical this model ends up being, it’s tough for teams to truly understand the value of their leads. Sure, a +75 score is better than +50 but often there’s little clarity over what makes for an okay lead versus a good one. The challenge gets that much greater when we start communicating cross-functionally.”
Ultimately, “traditional SaaS lead scoring just doesn’t work,” says Sharkey. “Rather than aligning teams, it tends to lead to distrust between them. The old way was built long before the notions of Product Led Growth (PLG) and try-before-you-buy SaaS had been conceived, so it’s no surprise that today's sales and marketing teams are frustrated with the experience.”
Andre Yee, CEO and founder of Triblio, agrees with Sharkey, going so far as to refer to the concept of the MQL as a “failed experiment”. “The evidence is staring us straight in the face whether we like it or not. The conversion rate of marketing qualified leads to the sales opportunity pipeline has been declining for years,” he writes. “We need to ask ourselves why this is happening and rethink the very essence of how lead generation is done.”
The rise of the product-qualified lead and a modern approach to scoring
SaaS product-led growth has forever changed the way business buyers interact with SaaS sales teams - if they ever interact with them at all - and flipped the marketing funnel upside down in the process.
“In the world of marketing automation, lead scoring is the primary method that marketing teams use to monitor how prospects are nurtured and determine the right moment for SDR teams to qualify prospects. With a product-led strategy, a product-qualified lead(PQL) serves this purpose. It is essentially an activated user that has realized initial value within your SaaS product, and is showing enough interest to trigger your sales team to engage with that prospective customer.”
In SaaS businesses that operate on a freemium or free-trial model, it’s entirely possible that a sales team need never engage with the vast majority of future customers, with these users able to realize enough value from their in-product experience (reaching the prized “aha” moment) and convert to a paid subscription completely of their own volition.
For those users who haven’t yet realized the full value of a product, it is their in-product behavior that should be measured, rather than their engagement outside of it. In a product-led business, writes Bush, “a product-led marketing team asks, ‘How can we use our product as the #1 lead magnet?’ A product-led sales team asks, ‘How can we use the product to qualify our prospects for us? That way, we have conversations with people that already understand our value.”
Scoring intent based on in-product behavior
Bush suggests that, instead of using the old approach to lead scoring, go-to-market teams should build a new framework to SaaS lead scoring based on in-product behavior and firmographic and demographic data that identifies PQLs. They can start by looking for patterns amongst their best and worst customers when analyzing their data and scoring these accordingly.
But it’s not enough to simply score these metrics. To be of any value to sales and marketing teams, they must be normalized to allow for comparisons and, furthermore, they cannot be static. While the demographic and firmographic elements of a user’s score might remain the same, the behavioral metrics should not. This tends to be the point at which sales and marketing teams in product-led businesses hit a roadblock and revert to the model with which they are most familiar - even when they know the value of in-product scoring would lead to better outcomes.
“We now talk about Product Qualified Leads (PQLs), Customer Health Scores, Likelihoods and a whole bunch of other scores and scoring acronyms to help us make sense of fit and behavior at every stage of the customer journey,” says Sharkey. “But, up until now, the majority of us have just talked about these scores instead of actually implementing them – and for good reason. Building unique scores that degrade over time for behavior and fixed scores for demographics has been a complex and complicated process. You either need to involve developers and build custom scores, or evaluate a dedicated lead scoring product.”
Even then, “lead scoring is not a one-and-done activity,” notes Campbell. “To keep up with the constantly changing business environment, organizations must reevaluate and update their lead scoring models continuously. This is an impossible task to do manually, as the time, energy, and focus needed for this devotion far exceeds bandwidth and resources.”
Machine learning and AI in the development of lead scores
Enter machine learning. “Leveraging machine learning for AI-driven model selection takes the black art out of developing effective scoring models,” writes Campbell. “AI can continuously improve the lead scoring model based on actual performance. This provides an objective, measurable, and agreed-upon qualification standard for passing leads from marketing to sales, as well as the means to prioritize them and ensure the right sales and marketing resources are focused on opportunities that hold the greatest promise.”
It’s important to note, however, that Lead scoring using machine learning and AI will only ever be as good as the data you feed it. “ The challenge with embracing AI technology is that it's dependent on CRM data quantity and quality, which, as our survey revealed, is lacking for most companies,” notes OpenRise’s report. “The answer lies in enriching incomplete CRM data with additional attributes to improve AI learning models' efficacy. Data enrichment also enables AI to consider factors not present in CRM data, surfacing additional competitive insights previously unknown to organizations.”
Why businesses taking a product-led approach to SaaS lead scoring are those that will win
Product-led businesses are, quite simply, the future of SaaS. “The most profitable software companies with the best Rule of 40 are still PLG companies,” notes TechCrunch. “Products — rather than just people — are responsible for acquiring, converting, and expanding customers.” SaaS businesses taking a product-led approach to sales - and therefore to lead scoring - will gain efficiencies that go beyond acquisition to retention, allowing them to scale faster, allowing them to better deploy their resources and, ultimately, gain a competitive advantage.
From a go-to-market perspective, “a free trial or freemium model opens up your funnel to people earlier in the customer journey,” writes Bush. “This is powerful because, instead of prospects filling out your competitor’s demo requests, they’re evaluating your product.”
Many of today’s fastest-growing SaaS companies have implemented a product-led sales strategy, including Slack, Zoom, Asana, InVision, Expensify, and Dropbox, and sped up their growth in the process. “These companies realized that nothing is more valuable than understanding how customers use, interact with, and feel about their products,” writes the Aptrinsic team. “Not even NPS and customer satisfaction surveys can provide as much insight as actually monitoring and analyzing how customers interact with your product and service in the real world.”
There are numerous reasons why these businesses are able to grow faster than those still taking a sales or marketing-led approach to acquisition. For a start, as noted earlier, many prospects convert to paid without ever needing to speak to a salesperson, having already experienced the value of the product. For those that haven’t yet converted to a paid plan, go-to-market teams can focus their efforts on high-value prospects that match their ideal customer profile, whom their new lead scoring model identifies as being most likely to convert, instead of spending their time on leads who haven’t demonstrated any intent to buy and are unlikely to ever become customers.
For those prospects that do engage with the sales team, they’re able to abandon the “high-touch, sales-led approach to developing leads[which] relied on customer meetings, demos, proof-of-concept deployments, and the like to convert prospects to customers,” says Mickey Alon, Chief Product & Technology Officer at VidMob, because they are speaking to people who have already experienced the product. They can therefore leverage “strong and clear product usage signals associated with demographics and firmographic data [to] deliver initial value to those prospects and close deals faster.”
The Aptrinsic team agrees. “The insight that companies can extract by analyzing how prospects and customers use their product is by far the most important element of building a highly successful company,” they write.
How a product-led approach to lead scoring aids retention
It’s not just the sales team who benefits, either. The customer success team can better use their own time too and spend their time focusing on those customers who are not only the most likely to renew or upgrade but also those who are most likely to churn. “Once you measure product engagement, you can sniff out activity churn and combat it before it happens,” writes Bush. “Unlike customer and revenue churn—which look in the rear-view mirror—activity churn looks ahead and can save accounts before it’s too late.”
The Aptrinsic team agrees that this method is by far the most effective retention strategy. “We are not advocating against the use of NPS and customer surveys. We are simply arguing that product behavioral data uncover more insights faster, and is the cornerstone of delivering a great customer experience.”
Bush believes that, as the SaaS industry continues to evolve, there will be two types of companies: “Sales-led companies represent the old way. It’s complex, unnecessary, expensive, and all about telling consumers how the product will benefit them. These companies want to take you from Point A to Point B in their sales cycle.”
Ultimately, write the Aptrinsic team, this new approach “empowers SaaS companies to better predict prospects with the highest CLV potential, and take actions with customers to increase the likelihood of achieving that goal.”
“Simply put,” they conclude, “adding predictability to the equation enables SaaS companies to take their businesses to new heights faster than ever before.”
Implementing a product-led SaaS lead-scoring framework
As more and more businesses become product-led, implementing a new way to lead score will be crucial to uncovering and retaining high-value prospects. That process begins and ends with data.
Enabling “granular visibility” of product behavior
“Usage data is the foundational building block of any product-led motion. Usage provides the intelligence that drives all other functions, from pricing and packaging to sales, support engagements, and even product roadmap development,” the CEO and co-founder of Amberflo.io and former general manager at AWS continues. “This data shows which features are driving traffic and adoption."
Connecting the dots
Data on in-product behavior is not enough to build out an effective lead-scoring model on its own: this data must be combined with firmographic and demographic behavior in order to surface high-value prospects. This is where customer data platforms (CDPs), which bring data from otherwise siloed sources together in one place - like your product, CRM, and marketing platform, for example - come in.
“Uncovering the right leads is one of the key challenges of freemium models: Some customers will never convert out of the free tier, while others could bring very valuable revenue into the fold, as long as they get pitched the right offer at the right time,” continues Haim. “But knowing who’s who requires connecting the dots between product usage and the tools that marketing and sales teams use in their day-to-day.”
Building a product-led lead scoring system
Once the data infrastructure is in place, a lead-scoring approach based on product usage data can be implemented. These metrics will look different for every business, writes Moritz Dausinger in his guide to building out a lead-scoring framework to identify product-qualified leads. “For some brands, even logging into the app regularly or performing specific tasks will signify purchase intent. For others, it could be inviting colleagues to the app or visiting the pricing page a couple of times.” Remember that the scores for behavioral elements should degrade over time - a user who logged into your platform yesterday shouldn’t have the same score as one who hasn’t logged in for a year - while the demographic and firmographic elements will remain fixed.
Bush suggests answering the following questions to identify the set of metrics or actions that align with your business goals, which you can then use to build out a lead-scoring framework:
What do your best customers do regularly in the product?
What do your best customers not do in the product?
What features did your best customers try first during onboarding?
What are the similarities among your best users—demographics, team structure, ability?
It’s equally important to identify the behaviors that indicate which customers aren’t the right fit for your product, which is why Bush also suggests looking at patterns in your churned customers:
What were some of the main differences between their user journey and that of your best customers?
What outcomes did your churned users achieve and not achieve?
It’s crucial that businesses continue to evolve these metrics as they gather more data, which is where AI and machine learning will also play a key role. “Over time,” notes Gupta, “you should begin to uncover more sophisticated insights from the usage pipeline, such as checkpoints on the typical customer onboarding journey from signup, or identifying the important levers for activating, increasing and retaining users on the platform.”
Dausinger says that, while “building the PQL model seems intimidating when you think of it first, in reality, it is actually quite easy. The secret to it is to know the characteristics and actions that signify product activation and using them to score leads and uncover the biggest sales opportunities.”
Andrea Warmington is a content strategist and writer, who has been working in content for 10+ years. She started her career as a journalist before moving into the world of content strategy, for both B2B and B2C businesses. She has a lifelong love of storytelling and believes in taking a journalistic approach to all of the content she creates. In recent years, she's developed a real passion for leading transformative content projects that establish tech businesses as thought leaders and reputable publications in their own right.