First time I came across lead scoring, I thought it was total b.s.
A random number calculator, baked into your marketing automation tool of choice, with no bearing on reality. Whoopee… we just got a 63!
It seems like a joke. But sales and marketing alignment isn’t. Teams own tools; tools own data. When you truly integrate all your tools and data, you also integrate your teams. Sales and marketing have to work off the same data.
Through working with customers at Hull and observing data-driven teams, we’ve seen this transformation and realignment happen firsthand. And with that, the transformation of practices like lead scoring.
And it turns out, lead scoring is actually incredibly useful.
You just need to pivot you’re thinking about what it’s useful for…
Lead scoring: A useful concept. A useless output
Sales-marketing misalignment starts with misaligned goals. Lead scoring is all about qualifying and nurturing leads; sharing a common goal and hand-off.
The most common drive of sales-marketing misalignment we’ve observed is a mismatched lead qualification process. Marketing generates leads and lobs them over the fence to sales. Sales then have to follow up and close the deal. Except sometimes marketing’s leads suck. Or sales doesn’t follow up properly.
In theory, attaching a score to assign the value to a lead should help reps prioritize. But in reality, this doesn’t happen. Reps have their own ideas and biases. All the detail is obscured and presented in some seemingly random lead score. What’s the difference really between an 80 and a 90? How did they get there? What should I say to this person?
The output is useless… to the rep!
What’s the real goal? The job-to-be-done?
Tinder isn’t about accumulating matches. It’s about starting conversations that lead to dates.
Ebay isn’t about accumulating bids. It’s about winning auctions that get you the goods.
Lead scoring isn’t about qualifying leads. It’s about empowering sales reps to take action.
Josh Hill from discusses lead scoring here -
Lead Scoring is touted as a way to help Sales prioritize leads, but is often used by Marketing to pre-qualify leads with the assumption that higher scores equals higher priority. I do not think that is logical because the sales process is more complex than that.
Josh HillMarketing Rockstar Guides
There’s a difference between the activity that’s measured (the score) and the action that needs to happen (the sales reps job). At Hull, we think of sales enablement as three goals:
- WHO to talk to (qualified leads, ready for sales)
- WHAT to say (the context & conversation starters)
- WHEN to say it (the timing and notification when action is needed)
Given those three jobs-to-be-done, where does lead scoring fit?
Lead scoring does not work “out of the box”
If you think of the “goal” of lead scoring as empowering sales reps instead of qualifying leads, then you can see how turnkey lead scoring doesn’t work. Your business is unique. So are your leads. And so is your sales process.
First, scoring needs to inform who to talk to. Numbers can still help prioritize, but this has to match up with what is really going to drive success for your sales reps role, not just the level of marketing activity of a lead.
Second, your score needs to inform what to say. WHY are they a qualified lead? You need to pair quantitative scoring with qualitative signaling.
Third, you need to integrate scoring into your other data and workflows so sales and marketing use scoring to take action at the right time.
All three of these tasks will be unique to your company. And this is how we see data-driven teams solve for lead scoring.
Tactical Takeaway #1: First, identify fit amongst your leads
First, you need to identify best-fit potential customers amongst all your leads.
These aren’t based on arbitrary scores — 8oz’s of sand weighs as much as 8oz’s of steak — but whether any leads fit your clear, common, objective definition of your ideal customer profile.
Your ICP will define the data you need to source and shop for. For the data you need to identify your ICPs that you can’t source 1st party from leads (e.g. through form fills) or infer from their behavior (like selecting categories on your website or in-app), you need to use data enrichment to fill-in-the-blanks.
Wherever your leads are coming from, you need to be able to enrich them all and build that complete picture so you can accurately qualify (and disqualify) them all. Then, you can use absolute rules to measure fit — they either match your criteria or they don’t. A score isn’t helpful if they’re in a location, language, industry, or size where you can’t service them.
You still need to have a segmentation engine that can build these precise segments from many different sources, and use that to query your entire customer database.
Voila! You know have identified whether a lead is a good fit or not.
Tactical Takeaway #2: Score every “aha!” moment from universal lead tracking
A measure of fit is not enough to fully enable sales to talk to the right person. You need to know how engaged those best fit leads are — whether they’re cold or “hot”.
But leads interact with your brands in more places than ever…
- Live chat & chat bots
- Review websites
- Free trials in your product
- Multiple owned websites
- Every kind of social media
… and so on. The danger is there’s no unified lead profile and understanding of how truly engaged a lead is with your brand.
This is where we see teams use universal lead tracking — tracking every interaction and lead source across every channel.
With every moment in the customer journey being tracked into one place, you can spot the “aha!” moments when people move through the buyer’s journey. There are three methods we spotted data-driven teams using to do this:
- Raw product usage (a “guess” — good enough to get started)
- Quick regression analysis
- Predictive lead scoring (more on that later…)
In general, we see data-driven teams starting from progressing down this list (vs. jumping straight in with the data scientists). It’s an iterative process, not a fire-and-forget project.
The goal is to understand which behaviors (of all a lead’s behavior) correlates highly with purchasing and “healthy” product usage. These moments - the full profile of your leads engagement and the moments that matter - are then captured in your lead scores.
Voila! You can now build a score based on every leads engagement across every channel
Tactical Takeaway #3: Consolidate complexity of leads interactions into lead scores
The problem with universal lead tracking and ideal customer profiles is they involve a lot of data and a lot of different types of data. And even though this is helpful to qualifying the right leads, it can be overwhelming for a rep and cumbersome to manage.
Different types of data are not like for like. Different page views have different values (a blog post vs your pricing page vs. your enterprise pricing FAQ page). Page views have different values to live chat conversations have different values to form fills have different values to in-app product usage…
This gets messy.
Instead of building bajillions of complex “rules-based segments” to capture everything, we notice data-driven teams use lead scoring to capture all that complexity into one simple trait.
Numbers can feel arbitrary, and leave room for interpretation (particularly when calculated with data from many sources). Instead, choose a measure with fewer discrete units. For instance, lead grading (A, B, C,…) or using phrases to describe fit (“Very good”, “good”, “medium”, “low”).
This creates clearer buckets which you can separate out in an automated workflow or task list for a sales rep. You don’t want reps guessing and dithering when it comes to your Grade A, best-fit leads!
The final piece of complexity is the temporality of leads. Leads who request demos, attend webinars, and start chat conversations need to be followed up with quickly (hours, maybe even minutes…) to capture that interest whilst it’s still hot.
We notice the best marketing teams stipulate follow up times with their sales teams in an SLA to make sure this happens. (Subscribe to our future “Spotted” post on sales and marketing alignment).
For lead scoring, you need the recency of actions taken to reflect in their score because the leads aren’t as “hot” after time has passed. Your score needs to decay (with a half-life function). This means the lead scores should always be accurate at any point in time in telling the sales reps who to focus on.
Here’s an example what this equation looks like for lead scoring with decay.
Voila! You’ve collapsed all the complexity of your leads into a simple-to-understand score.
Tactical Takeaway #4: Give sales a conversation starter, not just a score
Remember the original job-to-be-done. Lead scores aren’t about qualifying leads, but about enabling sales reps. The most important thing to give here is a conversation starter, as John Sherer says.
[Sales] doesn’t want this intelligent thing that says “this lead is 66% more likely to buy”, because they can’t use that to communicate with the prospect. […] But they can reach out to a prospect and say “hey, you just installed. What are you looking to do next?
John ShererDirector of Sales, Appcues
Now imagine you’re sitting with your next lead, perhaps over a coffee or a beer. What clues and cues would you find really useful to make conversation and point them towards seeing the value in your solution?
What information about their company? Themselves? Their recent behavior?
For a sales rep, you’ve got to help them join the dots. This (brilliant) sketch from the @GapingVoid helps illustrate this idea…
But as John says, “sales doesn’t want this intelligent thing” to tell them exactly what to do. It’s not a prescription, just direction. With the coordinates set, let the sales reps hone in and engage human-to-human.
They need the conversation starters. These are called signals.
Signals set the context for the conversation. They can use to cue in sales reps, control automated workflows, and personalize templated content.
So how do you package your data up? What sort of insights are useful? (The best teams will pull these from discussions with sales reps, and observing the sorts of discussions and searches they do already, but…) here are a few ideas.
Highlight existing data
Most data-driven teams have dozens upon dozens of attributes per person and per company. This becomes difficult to wade through at a glance. Highlight the key talking points.
- Uses Salesforce (instead of displaying all their tools)
- Referring domains (instead of displaying every new session)
- Not English speaking (instead of parsing through all their messages and location)
Highlight existing trends
Summarize streams of events (which would take forever to dig through for a sales rep) into useful talking points.
- Visited pricing page 3+ times this week (instead of showing every page view ever)
- Most interested in “personalization” (instead of showing every blog post read ever)
- Invited 5 users in their trial (instead of showing every single event)
Recency in general, and of specific actions can give some context as to how often
- Last seen (instead of showing all events)
- Last login (instead of showing all events)
- Last seen on the blog (instead of showing all page views)
To create signals like this to match your sales reps needs, you need to be able to run custom computation and data transformation on all your lead and customer data. This is often a challenge in traditional marketing automation platforms (which usually host lead scoring algorithms), particularly when combining multiple types of data (marketing engagement, in-app product usage, chat conversations, web analytics, social media engagement etc.).
To solve this, we see data-driven marketing teams use separate scripts, customer data platforms, or dedicated lead scoring tools to manage all the complexity of computing signals.
Importantly, all signals and scores need to be synced to your CRM and sales enablement tools in as close to real-time as possible so sales reps can take action with it right away.
Voila! You can now leverage your lead and customer data to personalize your sales engagement.
Tactical Takeaway #5: Use predictive scoring models and tooling
The problem with custom computation (scoring, signals, transforming data to sync to your CRM and key marketing tools) is it’s difficult. Legacy marketing automation platforms aren’t set up for this vast and complex modern customer data management we see data-driven teams designing.
The first challenge is the inability to flexibly transform data. For instance, to increment (and decay) a lead score according to dozens of data points tracked from different sources, or to compute meaningful signals (contact and company level traits) to inform a sales rep.
The second challenge is simply managing the complexity. In the second tactical takeaway on scoring every “aha!” moment, we shared the three methods people use. Guessing, regression, and then predictive score. The first two involve “rules-based” logic. “Rules-based” logic falls apart when managing large data sets. If this then that rules have to be defined by a human and are set in stone, but who says the human is "right"? And continues to be right?
Humans might be great at communicating and empathy, but analyzing large data sets is not best solved by a human. Can you spot trends across 10,000 profiles with 100 traits each? To “spot” more opportunities amongst their existing leads, we notice teams turning to lead scoring tools.
The hardest part remains in rendering the data actionable. This is where big data can help personalization at scale. Lead scoring tools have been built with this in mind. They leverage the multitude of data points available to automate the qualification historically run by SDRs.
Francis BreroCRO & Co-Founder at Madkudu
The choice tool we see teams in B2B SaaS use for lead scoring is Madkudu. They build a custom, predictive model on top of your existing data and identify the “aha!” moments you’d otherwise miss. They do this by unifying data, deriving computed traits from that data, then training predictive models on that data to find the best measures of fit. (Here’s an overview how it works).
One key advantage of artificial intelligence and machine learning is the ability to iterate and update their models. This enables you to evolve and optimize your models and qualification as new leads come in.
It’s important however to consider predictive lead scoring as optimization, not as a source of new leads. Typically, we see lead scoring and lead scoring tools like Madkudu inserted into existing end-to-end data flows (for instance, from leads marketing’s marketing automation platform to sales’ CRM) which are already delivering results.
Voila! You can now leverage predictive data to “spot” more deals amongst your leads
More "Spotted" playbooks like this?
Subscribe to The Crow's Nest newsletter from Hull for best practices, ideas & resources for customer data management.
Tactical Takeaway #6: Transform lead scoring into action
So you have a lead score. Now what?
One consistent pattern we notice data-driven teams follow is they use data for action, not just insight.
Lead scoring should be at the heart of your lead qualification, marketing-sales handoff, and sales rep notifications. That means you need to integrate your customer data so it changes you the workflow of in your marketing automation and your sales reps workdays.
To quickly recap the first parts of our five-part personalization framework and how it impacts sales enablement:
- WHO are you messaging
- WHAT are you saying
- WHEN are you saying it
Lead scoring needs to inform all of these.
High and low scores (and different ranges of signals) should control the segments and workflows that your leads enter. For instance, you might assign very good fit customers to a sales rep, and send an automated email to follow up with their calendar link to book a meeting right away. And enter that company into a retargeting strategy to try to catch more meetings elsewhere.
Next, your lead scores and signals should define what message people receive. This is partly controlled by the segment (they’re going to receive emails A, B & C), but also by dynamic content. Using different signals, you can start to swap out words, sentences, paragraphs, images, and more across all your messaging.
For instance, at Hull, we compute a
primary CRM trait for each company lead from the array of technologies from Datanyze. The talk track and messaging if your primary CRM should be (and is) different between Salesforce users and HubSpot CRM users.
Most importantly, you need to be able to act and react in real-time. Leads and leads behavior (which is being scored) are temporal. The time to act on a high scoring lead is NOW. With marketing automation, this can be triggered automatically by entering a segment or workflow, but some of the key engagement comes from 1:1 manual outreach. To trigger this, we notice teams using email and Slack notification.
More tools used by data-driven marketers (like data enrichment and lead scoring) are providing native Slack integrations. Clearbit, Salesforce, Madkudu, and others offer direct Slack integrations to empower teams with customer data.
Slack democratizes data like it democratizes communication - by making the data easily accessible to anyone on your team, you’ll be able to align your teams to action easier.
(Many Slack integrations will let you query your customer database too, so it’s not just a one-way feed).
However, with multiple systems holding useful lead & customer data, your sales reps may end up with a confusing array of different notifications from different tools for every lead. Assignment notifications from your CRM, scoring notifications from a scoring tool, enrichment notifications and queries from another service…
The smartest notification systems we’ve spotted combine and condense multiple sources of data - everything that matters - (remember, the goal is to start a conversation) into a single notification. This gives sales a very high signal-to-noise channel of information.
We notices data-driven marketers use sales rep notifications to trigger, personalize, and iterate on sales conversations like they do for automated, ‘broadcast’ messaging (like email, live chat, ad audiences etc.).
We’ve noticed this kind of lead scoring & notification empowers sales reps to use their data more. Instead of relying on old school assumptions and heuristics, they spot patterns and “connect the dots” with each of their leads to reach out, talk, and create deals in a way they haven’t been able to before.
Now the appetite among the team to actually do this has changed dramatically because there's a realization that the messages can be very targeted and can be very helpful and don't have to appear to be marketing-y or sales-y.
Emmanelle SkalaFormer VP Sales & Customer Success at DigitalOcean
Voila! Now your sales reps are empowered to take action by their data.
Best-fit criteria for lead scoring and signals
Lead scoring is not for everyone. We’ve observed the best results amongst customers who fit these criteria:
1. You are post-Product-Market fit
You have a defined ideal customer profile that you can support today. From this, you have defined your messaging, channels strategy, pricing model and so on.
2. You have a steady volume of qualified leads already
From your PM fit, you’ve already at least one scalable, repeatable channel to drive a steady flow of leads that your sales reps can close in a reasonable time frame.
More (MOAR!) leads is always great, but you’re at the scale where a 10-20% increase in leads marked as qualified converts to a meaningful number of new deals.
3. You have a sales process that works
You can close the qualified leads you have. Your reps have messaging that works and an effective sales process nailed down. You’re seeing reps consistently hit quota (not just “superstars”). You’d double your sales team if you doubled the number of qualified leads.
Results we’ve seen
There are two trends we’ve seen:
- An increase in qualified leads and sales opportunities (varies from company and circumstance, but up to 4X increase in qualified leads, and tripling of sales opportunities)
- Reallocation of time amongst sales reps. They’re able to speak to right people with the right message at the right time.
Did you learn something new?
We at Hull research how to "join the dots" between your tools, teams and data. Subscribe to follow along with more articles like this, and learn the latest trends, tactics, and techniques.
What you should do now
If you'd like to learn our best practice for customer data integration, read our free Guide to Getting Started with Customer Data Integration.
If you enjoyed this article, perhaps your team will too? Why not share it with the links below.
Get started with data enrichment & lead scoring
Fill in the blanks in your lead & customer profiles, and compute lead scores via Hull's integrations.