Segmentation Tab

Segmentation helps you identify your best future customers by learning from your existing ones.

Instead of manually applying filters like “industry = SaaS” or “company size = 50–200”, you provide a list of your best customers. Segmentation then analyzes what they have in common and finds similar companies.

In simple terms:
You show us your best customers → we find more companies like them.

Segmentation uses Ocean’s database of tens of millions of companies and analyzes far more than basic firmographic data.

It evaluates signals such as:

Based on your input, Segmentation builds a customer profile model and ranks other companies by similarity.

👉 This goes beyond traditional “lookalike” tools that rely only on filters like industry or headcount.

Segmentation is designed for:

Segmentation is currently available in early access.

To enable it:

Go to app.ocean.io and select Segmentation from the left-hand menu.

Click Get started, then choose one of the following:

Select how specific your segments should be:

👉 Broader = fewer, larger groups
👉 Narrower = more specific, smaller groups

Choose which data points Segmentation should use:

If no changes are made, Ocean automatically selects the most relevant patterns.

Click Run segmentation.

⏱️ Processing typically takes 1–3 minutes, depending on list size.

At the top of the Segmentation window, you’ll see a summary of your segmentation results, including:

This gives you a quick overview of how your segmentation was built and what data it is based on.

Below the overview section, you can:

👉 This helps you quickly navigate and organize your results, especially when working with large datasets.

Your companies will be grouped into Segments, where each segment represents a different customer type.

By clicking on any segment, a detailed view will open in a new window where you can:

Each company in a segment receives a similarity score that indicates how closely it matches the selected segment.

The score is based on:

👉 The more a company aligns with high-weighted values, the higher its score will be.

If a company scores highly across multiple categories, it will appear higher in the ranked list.

Similarity scores range from -100 to 100:

👉 Higher scores indicate a better fit to the selected segment criteria.

If you are segmenting based on company size, the scoring might work like this:

👉 In short: the closer a company is to the ideal target range, the higher its score will be.

All selected values and category weights can be adjusted at any time:

This allows you to continuously improve the accuracy of your ideal customer profile.

Inside the segment detail window, you can switch between different tabs:

This helps you understand:

You can also click the Show lookalikes button on the right of each Segmentation to run a completely new search based on your segmentation. This allows you to further refine the results by adding or adjusting filters as needed.

From each cluster, you can:

From the Segmentation dropdown menu on the left hand bar, you can access all previously created segmentations.

Here you can:

At the top of this window, you can:

👉 This allows you to easily manage multiple segmentation projects and start new ones without losing your previous work.

Most tools rely on simple attributes like:

Segmentation goes deeper by analyzing:

👉 This leads to more accurate ICPs and higher-quality lead lists.

Each segment will include a plain-language summary, for example:

“Fast-growing SaaS companies in marketing automation with strong hiring in sales roles.”

You’ll be able to see:

This helps identify your most valuable segments—not just the most common.

The model is trained on your uploaded companies. It identifies shared traits (e.g. industry, growth signals, business characteristics) and uses them to rank similar companies.

Yes, with some limitations:

No. Your data is private and used only to build your segmentation model. It is not shared or used to train global models.