HubSpot deal scoring: a complete guide for sales teams

Sales teams often struggle to know which deals will actually close. You might have a pipeline full of opportunities, but without a clear way to rank them, your reps spend time on the wrong accounts. This leads to missed quotas and inaccurate forecasts.

HubSpot deal scoring solves this by assigning a value to every deal in your CRM. It helps you see which prospects are ready to buy and which ones are just kicking tires. In this guide, we will walk through how to set up deal scoring and why moving from manual rules to AI is the best way to grow your revenue.

What is HubSpot deal scoring?

Deal scoring is a method used to rank the sales readiness of a deal. In HubSpot, this is typically a number between 0 and 100. A higher score means the deal is more likely to close.

There are two main ways to handle this in HubSpot. First, there is manual scoring. This is where you set up rules based on deal properties. For example, you might add 10 points if the deal amount is over $5,000 or subtract 20 points if the deal has been in the same stage for more than a month.

Second, there is AI deal scoring. This uses machine learning to look at your historical data and predict the outcome. HubSpot is currently updating its infrastructure. According to HubSpot's 2025 Product Update, legacy score properties will stop updating on August 31, 2025. They are moving toward a more integrated AI ecosystem.

How to set up deal scoring in HubSpot

If you are using manual scoring, you need to define the criteria that matter most to your business. Here is how to get started:

  1. Identify your key properties. Look at your past wins. Did they all have a specific industry in common? Did they all have a decision-maker involved early?
  2. Create a score property. Go to your HubSpot settings and create a new property with the "Score" field type.
  3. Add positive and negative attributes. Set up rules. If a deal has a "Demo Completed" status, add points. If the "Last Contacted" date was more than 14 days ago, remove points.
  4. Test the score. Use the "Test a Record" feature to see how a specific deal is scored. This helps you ensure your rules are working as expected.

Manual scoring is better than nothing, but it has flaws. It relies on your gut feeling about what matters. It also does not update automatically as market conditions change.

Why AI deal scoring is more accurate

AI deal scoring removes the guesswork. Instead of you deciding that a "VP" title is worth 10 points, the AI looks at thousands of past deals to see if that title actually correlates with winning.

Research by Rezazadeh (2020) found that machine learning models provide significantly higher accuracy in B2B sales forecasting than traditional human evaluations. The study showed that data-driven workflows help teams focus on deals with the highest monetary value.

Another study by Yan et al. (2015) highlighted that moving away from subjective human ratings to predictive analytics allows companies to capture the real influence of seller activities. The AI can see patterns that a human manager might miss, such as how the frequency of CRM updates relates to the final win rate.

AI models are also dynamic. They retrain themselves as you close more deals. This means your scoring gets smarter over time without you having to manually adjust any rules.

Using deal scores to prioritize your day

Once you have deal scores in HubSpot, you should use them to change how your team works. Here are three ways to use these scores:

1. Focus on high-probability deals

Your reps should start their day with deals that have a high win probability but are still in early stages. These are your best opportunities to hit your monthly target.

2. Identify at-risk deals

Look for deals with high values but low scores. These are "at-risk" deals. They might be stalled because of a lack of engagement or a missing decision-maker. By spotting these early, you can step in to help the rep save the deal.

3. Improve forecast accuracy

Instead of asking reps "How do you feel about this deal?", look at the score. If a rep says a deal is 90% likely to close but the AI score is 20%, you have a coaching opportunity. This leads to a much more realistic sales forecast for the whole company.

How Aigenture automates deal scoring

While HubSpot offers basic AI scoring, it is often "generic by design." It uses broad patterns that apply to many companies but might not fit your specific sales cycle.

Aigenture takes a different approach. We build a custom machine learning model for every single customer. We train the model on your historical HubSpot data. This means the win probability scores you see are based on how your customers actually buy, not a global average.

Aigenture installs directly into your HubSpot CRM. You can see the win probability and deal health insights right on the deal record. We also provide a "What-If Simulator" that lets you see how changing a deal amount or stage would affect the score before you make the change.

If you want to see which deals will actually close this month, you can View our Plans or start a 14-day free trial.

References

  • Rezazadeh, A. (2020). "A Generalized Flow for B2B Sales Predictive Modeling: An Azure Machine-Learning Approach." Forecasting. Link
  • Yan, J., Zhang, C., Zha, H., et al. (2015). "On Machine Learning towards Predictive Sales Pipeline Analytics." Proceedings of the AAAI Conference on Artificial Intelligence. Link
  • "HubSpot's New Lead Scoring: Your Guide to the August 2025 Update." HubSpot. Link