Conversational Chatbots for a More Accurate Lead Scoring System

Illustrator: Jana Pérez
lead scoring system

Not all leads were created equal. That’s why you are in sore need of a chatbot lead scoring system!Well, a chatbot-driven lead scoring system to be precise.

To manage your time and resources properly you need to know who is just browsing around; who is actively looking for a solution to their problem and who is on the verge of buying your product.

It doesn’t make sense to give them all the same amount of time, energy, and attention. You should focus your resources on people who are most likely to convert into paying customers.

But how can you know which leads are the most valuable?

That’s where lead scoring matrix comes into play!

What is Lead Scoring?

Lead scoring refers to assigning points (a value) to leads based on how likely they are to purchase your product.

Basically, you take a look at your customer base and try to identify what they have in common.

For example, let’s say you are selling accounting software, and you offer freelance, small business, and enterprise packages. You might find that freelancers are often unwilling to pay for it, meanwhile, large companies want a more sophisticated solution than the one you offer. That leaves you with small businesses as your most valuable leads since they are most likely to convert.

Knowing this allows you to prioritize and direct the majority of your resources to this specific segment of your target audience and hence increase your overall ROI.

Of course, this example is very simplistic.

Real-life lead scoring often involves developing a complicated system that takes multiple data into account to assign a score from 1 to 100 points.

Still, while the granularity might differ, the essence of what is lead scoring stays the same: using data to decide which leads to prioritize.

Why is it Important?

Lead scoring is crucial to allocating your resources wisely. It’s key to applying the efforts of your marketing and sales teams more efficiently and creating a more effective sales process.

resource allocation b2b marketers

Building on the example from above, if B2B marketers treat freelancers, small businesses, and large companies as being equally valuable, they will end up spending too many resources on the leads that are the least likely to convert and too few resources on the leads that are the most likely to convert.Furthermore, a good lead scoring system allows you to streamline your lead management, since tailoring your approach to each potential customer makes it much easier to get your leads to progress to the next step of the sales funnel.

For example, if you know that someone is a freelancer, you can leverage content marketing efforts. Send them a case study that demonstrates how someone just like them used your accounting software to get their finances under control and emphasize the specific issues that freelancers have to deal with.

Meanwhile, if you know that someone is a small business owner, you can send them a case study on how a fellow entrepreneur used your product to systematize and automate the company’s finances, and focus on the pain points specific to entrepreneurs.

Surely, you can see why this approach is more effective than treating freelancers, small businesses and large corporations exactly the same, and putting them all through a generic sales tunnel.

What Data to use for Lead Scoring?


In her article “Lead Scoring 101: How to Use Data to Calculate a Basic Lead Score” Hubspot’s Lindsay Kolowich suggests six different lead scoring models:

  • Demographic information. Age, gender, nationality, education, job title, interests, etc.
  • Company information. Industry, company size, annual revenue, B2C or B2B, etc.
  • Online behavior. The number of times someone has visited your website, which pages they visited, how much time they spent on those pages, etc.
  • Email engagement. Open rates, click through rates, email replies, etc.
  • Social engagement. Likes, shares, follows, retweets, comments, etc.
  • Spam detection. Does the information they provided look legit? Do the name and email they gave when they subscribed to your email list look like a real name and email?

Which data is the most important? This is THE question to answer as collecting ALL data can turn out to be more counterproductive than helpful.

It all depends on what you are selling and who is your ideal customer.

Most Common Lead Scoring Mistakes

Before moving to see how to create a lead scoring system tailored to your business needs, it’s important to highlight the most common traps and consider potential chatbot mistakes.

  • Failing to score your leads on both fit and interest.

“Fit” refers to lead characteristics and “Interest” describes lead behavior. Both of the factors are important, hence you need a lead scoring system with rules that cover both categories. Points should be spread more or less evenly between them.

Your most valuable leads will be those that fit the desired characteristics as well as exhibit desired behavior (e.g. repeatedly visiting your website, getting in touch for more info, downloading an e-book).

  • Scoring without structure.

Prospect scoring is no time for “guestimation.” Forget about “This is very important so let’s give it XY points.” Hence, it’s important to divide your points equally between several categories of rules, so you have a balanced point distribution across all influential factors.

  • Using an overly simplistic lead scoring system.

Simplicity can be the way in some cases but when it comes to lead scoring even subtle nuances can affect your chances. It’s not enough to divide leads into good and bad since such system doesn’t between good and best leads.

  • Not keeping your data database clean.

Unclean data encompass any information that is improperly formatted, duplicate, not-validated or out-of-date. A recent study confirmed that, on average, 25% of a b2b lead database is inaccurate and 60% of businesses reported their data health to be “unreliable.”

  • Thinking creating a lead rating matrix is a one-off job.

When you create your first scoring model, you are going to get some of it wrong as even the most balanced scoring will be, to a certain degree, intuitive. And that is OK. That being said, you should continue mining it for patterns and updating your approach.

  • Failing to create a lead nurturing structure.

If you have gone through all that trouble of creating a sophisticated scoring matrix, don’t waste it by not thinking of how to nurture the leads you qualified as most likely to convert.

How to Create a Lead Scoring System that Works for You?

scoring matrix that works for you

Here, we will lay out steps to create lead scoring rules that are complex enough to prioritize your leads effectively as well as provide you with a balanced point distribution that takes into account characteristics as well as behavior.

Questions to Ask:

  • What information do you need?

Before anything else, establish the minimum criteria any prospect must pass to become a customer. These are fixed qualification. If it’s not truly essential, don’t include it here.

  • Who is your target audience?

Next, identify usual but not necessarily essential qualities of your target audience/market. If you are unsure review the common attributes of your existing customer base. The idea is to identify prospects most similar to your typical customer.

  • Who is the perfect lead?

Now is time to define the characteristics of the absolutely “perfect” customer. Leads that are closest to this definition should earn a higher score.So, think about what makes certain leads better than others? What are their qualities and characteristics? Consider elements such as company size, budget size or position of your contact. Both marketing and sales team should give you input on this one!

  • What’s your customer base behavior?

After defining the characteristics, turn your attention to behavioral patterns. Make a list of all possible behaviors in which your leads can engage. Assign a point value to all behaviors based on which actions most closely correlate with your current customer base.Your list may feature web/landing page visits; number, and length of website sessions; contact requests; content downloads; form submissions; free trial activation; product demo requests/downloads; email opens; email replies/forwards; social media shares; etc.

Tasks to Undertake:

  • Pick a system.

Most businesses will do with the basic 1-100 point scoring scale. However, there are certain advantages and opportunities stemming from adding more digits to your scoring methodology.

In fact, it is particularly useful if you work with various types of leads. For instance, B2B lead scoring matrix can benefit from using the extra digit to differentiate between small, medium and large businesses by assigning them values of 1, 2 and 3. Hence, a prospect with a score of 3076 would refer to large business with a score of 76.

You could specify it further by assigning 100 points to each of the necessary criteria you identified in the first questions. This helps to weed out leads that are irrelevant but do exhibit high-valued behavior. For example, let’s say your prospect (to be a viable prospect) must be located in California, work in a specific type of industry (e.g., law or construction) and must have “manager” status in his or he respective company.

So, if a potential client only completes two of the compulsory criteria, it would have a score of 3276 (based on the previous example) or just 276 (if you are not using the fourth digit for differentiation). Despite exhibiting the right kind of behaviors (76) the prospect can be dismissed because it doesn’t fall into your customer base and would be a waste of time and resources.

  • Distribute points.

To prevent haphazard point distribution (as mentioned in the common mistakes sections) start by giving maximum value to your “categories of conditions.” Use it as a founding point for distributing the value to individual rules. This ensures you will create a more balanced lead scoring model.

For example, divide the points equally between characteristics conditions category and behavioral conditions category, fifty-fifty. Then assign values to the individual rules within those categories. This distribution will ensure that leads that display perfect behavior but lack necessary characteristics will never make it over 50 and hence waste your time and vice-versa.

  • Refine scores as you go.

It’s important to understand that establishing a lead scoring system is not a one-off project.In fact, this kind of attitude can result in analysis paralysis, because if you think you have to get everything right from the start, you might get caught up in trying to develop a scoring model so complex that it looks like it belongs on the blackboard in the movie “Beautiful Mind”.

You can run your first analysis after 4-5 weeks, you should already have enough data to work with. Run analysis and collect observations. Were there any low-scoring leads that ended up converting? Were there too many high-scoring leads that failed to convert?

Adjust your point system accordingly. When you collect more information, repeat.

How to Work out a Lead Score for Each Prospect?


Okay, but how exactly can you establish a lead scoring system, and work out a score to assign to each condition?

If you are not sure which behaviors or characteristics convert above your average, do the following:

  • Calculate the overall lead-to-customer conversion rate. In order to do that you need to take the number of new customers you acquired and divide it by the number of leads you generated.
  • Pick a condition that produced, you believe, higher quality leads.
  • Calculate the individual lead-to-customer conversion rate for that condition.
  • Compare that individual lead-to-customer conversion rate for that condition to the overall lead-to-customer conversion rate.

For instance, if your overall conversion rate is 1% but the conversion rate for leads that "requested a consultation" is 20%, then the conversion rate of the "requested a consultation" condition is 20X of your overall conversion rate. So, leads that meet this condition can be awarded, for example, 20 points.

You can start by focusing on a single characteristic/behavior and then identify other relevant attributes and include them in your lead scoring model.

The Surprising Benefits of Chatbot-Driven Lead Scoring


Your lead scoring model is only as good as your data. You can use analytics to gather quantitative data.However, when it comes to qualitative data, you need to reach out to customers. Most companies do it by asking clients to fill out various forms.

Of course, the problem with that is the fact that no one likes filling out forms as these feeble online tools go against the essence of the reigning experience economy. Not only is the response rate pitiful, asking people might alienate your clients.You have probably experienced this yourself. A website presents you with a form, you look at it, think “Later”, and leave.

How many times did you actually go back to complete that form?


Granted, if an obstacle like this deterred you from progressing to the next step in the sales funnel, you probably weren’t a hot lead anyway so an argument could be made that it was actually beneficial for the company in question.But while there’s some truth to it, ultimately it’s a naive view that doesn’t take into account user psychology, and subscribing to it might cost you a lot of money in the long run.

So what can you do?You can offer your potential customers a more enjoyable and interactive way to answer your questions: a conversational chatbot.

Practical Lead Scoring Chatbot Example

Home Genius is a business that helps people find reliable contractors and book estimates.

Take a look at the lead scoring chatbot below. The company uses a chatbot on its homepage to gauge the visitor’s intentions. By asking website visitors if they are ready to meet an estimator, they are already personalizing the conversation based on whether the visitor says YES or NO:

lead scoring chatbot example

If you click the YES option, you are prompted to specify what kind of estimator you are interested in (roofing, siding, solar, etc.).

lead scoring chatbot example

If you select the NO option you are met with an understanding message and a prompting to indicate when you might be ready (in 1 or more weeks).

You can take this approach to the next level by asking even more specific questions to qualify the prospective clientright there in the chat and then. With the best prospects you can take the next step down the sales funnel (e.g., to schedule a free 30-minute consultation at a time that is convenient for them) while the less "sale ready" prospects ones can be directed to helpful downloadable materials or blog posts.

Leverage the Potential of Conversational UI: Integrate Your Lead Scoring Chatbot on Multiple Channels

Of course, using an interactive conversational interface on your homepage is the most generic option. Hence, you can also enable similar interactive dialogues on your product, pricing, contact pages, a dedicated interactive landing page or make the bot part of a messaging app like Whatsapp.

And sure, you could get the exact same information by getting people to complete a form.

But really... if you were in their shoes, would you rather fill out a form or interact with a chatbot?I think it’s safe to say that most of us prefer the latter.

The other great application of conversational UI concept is creating chatbot feedback surveys you can send out to your existing customers to learn more about them and their need for and use of your product. Such insider knowledge can grant you crucial information to improving your marketing strategy, prospect scoring as well as the sales funnel.

Put the Qualified Leads through a Sales Funnel


There’s no point in scoring your leads if you don’t put them through a sales funnel.Leverage gathered email addresses and use an autoresponder sequence to educate them about the value of your product. You need to act fast as online leads have a very, very short life span.Ideally, these emails should be tailored specifically to them, as explained in the previous example of sending different case studies to freelancers, small business owners, and large corporations.

Also, including a clear call to action at the end of the introductory autoresponder sequence is absolutely crucial, you need to make your pitch and then tell the potential customer what to do next.Keep in mind that a call to action doesn’t necessarily mean asking them to purchase your product.

The aim is to get them to take the next step on your sales funnel. For example, you might ask them to sign up for a free 14-day trial, and then go for a close after their trial ends. You don’t need to try to get a sale straight away (a potential customer might be interested in your product, but not ready to buy it yet).

Generally, the more valuable the lead, the more resources you should spend on trying to convert them into a paying customer.

For example, you might:

  • limit your efforts to an email autoresponder for your least valuable leads;
  • offer a free trial or 30-minutes consultation to your most valuable ones.

Automate Simple Processes


To further your marketing automation you can use an online tool such as Calendly to handle your appointments. It’s an appointment scheduling software which allows you to create a calendar with available time slots.

You can then provide the potential customer with a link and they can pick a time that is most convenient for them. That way you schedule a call without having to go through email back-and-forth with them. Calendly even reminds them of the appointment so that they wouldn’t forget it.

Just make sure that everything you do is sustainable and that the ROI is positive (although doing things that don’t scale might make sense if you are just starting out and need to get your business off the ground).

Note: Using conversational interface for lead qualification allows you to qualify leads and react appropriately almost instantly. Unlike with an online form, you can set up your chatbot to recognize when a prospect meets your “best lead” conditions and arrange for a human takeover, schedule a meeting or facilitate a free trial.


It would be great if you could actively pursue each and every lead.However, the reality is that your resources are limited, so that simply isn’t feasible. You need to keep these constraints in mind and manage your time, energy, and money wisely.

That means focusing on the leads that are the most likely to convert into paying customers.

Lead scoring chatbot systems can help you easily identify who those people are & help you catch them while they are hot.

Are you ready?