AI-Powered ICP Analytics: From Customer Data to Revenue Growth

Featured Speakers:-

Links and Resources

• Lookalike Account Modeling for B2B Revenue Teams
• You’re Targeting Your Ideal Customer. So Why Aren’t They Buying?
• Adding the AI Advantage to Your B2B Lead and Account Scoring Maturity Framework

Transcript

[00:07] Taran Nandha

Hello, everyone, and welcome to this special session of MOPZA26, where we will be talking about ideal customer profiles and how AI is revolutionizing them.

The concept of ICP is not just about understanding who your ideal customer is, but also about activating hyperpersonalized campaigns around it.

With me, I have Arpit, who is the Founder and VP of Product Management at an AI-first platform for marketing operations experts. He also heads the teams for growth marketing and marketing analytics at Growth Natives, which is a full-stack, full-funnel agency.

I am also the Founder and CEO of both DiGGrowth and Growth Natives.

With that said, let’s move into our topic of the day: the ideal customer profile and what it should mean for you in this new AI world.

So, before we usually have this conversation with our customers, we ask a couple of questions, and I would love for you to also introspect and think about these two questions.

When was the last time you updated your ideal customer profile?

A lot of times, we hear people say that it was a year ago, or sometimes even longer than that. Rarely do we see people who say that it’s a more dynamic process and that they like to revisit it every quarter or every month.

The other question we typically ask is: How confident are you that the customers or ideal customers you’re identifying are the ones creating the most value for you?

What we have seen is that 20% of your customers drive 80% of the revenue, and an even smaller segment drives the most value.

So, there is a difference between revenue and value as well.

Just imagine that out of 100 customers, 20 may be the most important. And within those 20, there’s also a further classification you can do to identify who the real value customers are.

The reason is that people rarely ask questions about who’s likely to churn, who carries the highest lifetime value, which segment is growing the fastest for a company, and questions of that nature.

So, it’s not just limited to where we’re able to generate revenue, but also about where we’re able to generate the most profitable and easiest possible revenue.

An ideal customer profile should typically help you identify the path of least resistance when it comes to acquiring customers.

[03:21] Taran Nandha

We can go to the next slide.

Most of the time, we see companies limiting their ICP analytics to just a few factors, things like industry, revenue size, number of employees, region, and things of that nature.

Rarely do we see them really double-clicking on other aspects that can be far more valuable than just the basic demographics.

I’ll give you a perfect example.

If your product is not a good fit for prospects that are on a certain technology stack, then you should never be going after them.

So, if your product, for example, does not work with Microsoft Dynamics and the prospect base that you’re chasing uses Microsoft Dynamics, that itself is a flawed approach.

The more information you can bring into the mix when understanding the ideal customer profile, the better. There are hundreds of dimensions, and the beauty of this new-generation, AI-first approach to ICP analytics is that you can throw hundreds of dimensions at it, and it can still analyze them and identify the winning patterns you need to see.

At this point, Arpit, I think we should give everyone a sneak peek of the DiGGrowth Ideal Customer Profile Engine, and then we can come back to the deck if we need to.

[04:58] Taran Nandha

So, when we look at the Ideal Customer Profile Engine within DiGGrowth, it focuses on the foundation of any good ICP analytics: the quality of the data and how enriched it is before running the analysis.

That is the first thing we address in the Ideal Customer Profile Engine.

As soon as you plug DiGGrowth into your CRM, it is able to identify the gaps that exist in terms of your data completeness.

For example, it might show that 11% of your data is missing job titles, 19% is missing email addresses, or that you’re missing countries or other relevant information. It identifies those gaps very quickly, and within the system, you’re also able to enrich the data.

The moment you click on Enrich, it shows you all the records you have in your CRM. As you can see, there are certain dashes, which means that the data is incomplete.

Once you process the enrichment, all this data becomes complete, and you have a much better view of who your customers are.

[06:28] Taran Nandha

The next stage is running the ICP analysis.

Once you’ve run the ICP analysis, it gives you a complete picture of your customers and prospects that exist in your CRM today.

There are about 173,000 accounts in your database, of which 59,000 are customers. But more importantly, within that, only 20,000 fit your ideal customer profile.

This means that about 33% of your customers are driving about 80% of the revenue. It is also identifying the total revenue being driven, and 87% of it is being driven by just 33% of your customers.

On the extreme right, you can see that you have 82,000 prospects in the database. Just below that, it’s also highlighting that about 4,000 of them are Grade A accounts.

That’s where a lot of your focus should be: those 4,000 accounts. Your focus should also be on how to increase the volume of A-level accounts in your database.

[07:49] Taran Nandha

It also talks about the winning combinations.

In this dataset, the transaction size is small, which is why you see a lot of customers. But when you look at the winning combinations, it’s telling you what the best segments are.

As you can see, software and technology seem to be at the top, followed by business services and consulting. You can keep scrolling to see the other winning combinations.

[08:23] Arpit Srivastava

So, Taran, one quick question here. One thing that we generally hear from our clients is: What could be the minimum threshold for the number of customers where we should consider this kind of customer-based ICP analysis? Is there any direction there?

[08:39] Taran Nandha

Yeah, I would certainly say that you need to have a sizable customer base.

I would say that if you have anywhere north of 300 existing customers, this becomes really powerful. And as the number increases, it keeps getting more and more sophisticated.

So, if you have an existing customer base of about 300 accounts, this becomes very meaningful.

And then, this is really about the top segments. You can click on Industry Insights at the top, and you can see the top industries within that.

You can also go down to sub-industry levels and get both the customer and prospect lists right here, along with how many exist in each category.

This is showing you not just the numbers, but also the companies and, eventually, the people who matter within these accounts.

[09:45] Arpit Srivastava

Some of the industry insights here are going to be very helpful, especially for people in marketing who are running ABM campaigns.

That’s how you want to dissect the data. You want to understand which industries to focus on, such as software, and then run personalized campaigns against those segments.

Now, this is powered by evidence because you’ve already seen some success with these segments. You can leverage those insights further and apply them to your prospect database.

You also have a clear view of the whitespace available in a given industry or sub-industry.

[10:23] Taran Nandha

And then you can look at customer scoring and prospect scoring. You can also look at seasonality if that’s a factor.

[10:40] Arpit Srivastava

The scoring part is based on three dimensions.

The first is business insights. Any kind of business insights you share with us, including your strategy and prioritization, are grouped under the rule-based score. That’s the traditional approach we always start with.

But then it’s blended with a more data-driven and AI-driven score as well, so that you get the best of both worlds.

As you start activating campaigns, an engagement score is also added. This reflects how accounts and the contacts within those accounts are engaging and what their digital body language looks like.

[11:29] Taran Nandha

So, this shows you the level of depth we can go into.

One important distinction, as I was mentioning earlier, is that the number of dimensions you can analyze is almost unlimited.

The more information we can provide to the system, the more sophisticated and granular the analysis gets.

At the end of the day, you get the list of winning companies that should be on your radar.

Let’s move on to competitor analysis. That’s also a clear distinction for us.

Based on publicly available information, we’re able to identify the areas where your competitors are strong and where they are weak.

If we know that you’re stronger in certain areas compared to your competitors, you can run competitor displacement campaigns.

You can identify who the customers of those competitors are and which segments make the most sense for you.

That’s where you focus your energy when it comes to competitor displacement campaigns.

[12:47] Arpit Srivastava

Yeah, and behind the scenes, we leverage this data as well to influence the scores.

On one hand, you have your first-party CRM data, but these are more like market insights or insights about your competition: what their focus areas are, where they’re strong, or which additional industries you may have missed and could consider.

All that gets incorporated, and those insights are pretty helpful for the sales team when they’re reaching out to these accounts.

[13:22] Taran Nandha

And then we can talk about growth signals.

We’ve covered marketing: who you should be identifying and going after.

We’ve talked about sales: where they can focus their energy, which accounts they can prioritize, and how they should approach those accounts, including the individual people within them.

But growth signals are more about identifying where the most potential for growth exists within your existing customer base.

This is about upsells, cross-sells, total customer lifetime value, and the ability to renew.

All those things are predictively analyzed and scored so that you have a clear understanding of who is the right fit for your next product and who needs the most focus.

The nice thing about all this information is that it doesn’t just reside within DiGGrowth. It can be synced back into your CRM, your account management system, or your partner relationship management system—wherever the action is happening.

We surface the key insights, and that’s what the individual teams see.

For marketing, it’s really about: Who is your ideal customer? Who is your ideal prospect? And who are the people you need to go after?

Here’s your list. Now run campaigns against it.

When it comes to sales, it’s: Here’s a prospect who is most likely to convert and also has a very strong customer lifetime value. Go after them.

And then, for account management: These are our existing customers. This is the gold mine we can go after in terms of upsells, cross-sells, renewals, and all the other actions that need to happen.

[15:19] Arpit Srivastava

Yeah, and the idea is that we’re not just helping you with implementation and building the system. We’re also helping you operationalize it so that the whole revenue team is aligned and actually using these datasets to drive more growth and revenue.

And the most important thing is that it’s a living system.

It’s continuously getting enriched and driving insights in real time. The moment a prospect comes into your CRM, it tells you how likely they are to convert, how likely they are to become a customer, and everything else associated with them.

So, it’s a living and breathing ideal customer profile analytics system.

And, as I was saying earlier, this engine can also power activation.

One part of activation is that once you’ve identified the companies and people you want to go after, you can run digital marketing campaigns against them.

It could be paid advertising, social media advertising, or your sales team running individual campaigns against those target accounts.

And at a company level, once you have them in your prospect base, you can run hyperpersonalized marketing automation campaigns.

The DiGGrowth platform helps you create nurture emails that can be directly fed into your marketing automation platform.

And, Arpit, if you can demonstrate that, that would be great.

[17:09] Arpit Srivastava

Yeah. So, basically, behind the scenes, when we plug this in, we enter all the brand-related information.

Any kind of documents that you have—the system has all the context about the brand.

Then you have this magic prompt. Obviously, you can edit that.

Once you generate the email, based on whatever the system knows about the company, the contact, and any historical conversations, it generates a hyperpersonalized email copy.

So, this is just one example, and obviously, you can make some tweaks if you want.

We’re not owning the email engine side of it. We’re just owning the intelligence that is powered by your own first-party data.

All these email sequences will be plugged back into your preferred marketing automation system. It could be HubSpot.

From there, you can start any kind of workflow.

And then, as users engage with these campaigns, we get those signals back. That becomes the cycle Taran was mentioning: ongoing, always on, and always learning.

So, that’s how it works.

Likewise, we’re doing it for LinkedIn as well. The target accounts can be synced as an audience in LinkedIn, and from there, you can run ABM campaigns and LinkedIn ads against them.

[18:46] Taran Nandha

Okay, thank you, Arpit.

I think this brings us to the end of the demo.

We’ve covered most of the aspects that we had in our deck. We’ll just run through it again, Arpit, just in case we missed anything.

So, if you can move on to the next slide.

I think we’ve already talked about that, but it’s worth re-emphasizing that your existing customer data is extremely valuable.

If it is enriched properly, it can really point you towards who your next customer should be.

It’s also very important to do value ranking, which is identifying who your most valuable customers are, not just who’s likely to become a customer.

Then, you need to discover the patterns, create a common account score, and run it against your prospect database so that you can activate revenue against it.

And create a system that is continuously learning and always on, just like we demonstrated.

The Ideal Customer Profile Engine within DiGGrowth is always on, always learning, and always up to date with the best information to power your go-to-market campaigns around the best possible ideal customer profile.

A lot of times, when we’re doing AI-first projects, people are worried about the cost of running them.

We’ve done some value engineering on our side to make sure that we can keep the cost of tokens in check.

Arpit, do you mind describing that, because you’re closer to it than I am?

[20:32] Arpit Srivastava

Exactly.

We’ve actually started blending traditional large language models like Claude and ChatGPT with what we call the Jeff model.

This means that for decision-making, we’re not using generative models. We’re rather using the Jeff model.

And that helps us in two ways.

First, on the cost side, it optimizes token usage.

And then, in terms of getting inferences back, it’s a lot more optimized, so there’s no lag.

In that sense, we’re covering the best of both worlds: large language models as well as the Jeff model, which helps with better decision-making.

[21:18] Taran Nandha

Thank you, Arpit.

That brings us to the end of our conversation today. I hope you’ve learned something.

If you have any questions, please feel free to reach out to either Arpit or me, and we’ll be more than happy to jump on a call and walk you through a more in-depth showcase of our approach and how we do it.

And literally, the last thing I would say is that it may seem very complicated, but we’ve simplified it to a large extent.

It requires very little input from you as an organization or as individuals.

All we need is access to your brand guidelines, access to your CRM, and your prospect base, and we can run this analysis for you.

It typically takes anywhere between 5 and 15 working days to complete and give you a clear picture of what an ideal customer profile means for you and how you can activate demand generation against it.

Thank you, and I hope you have a wonderful rest of the conference. Looking forward to engaging with some of you in the future.

See more webinar