You can now use relevance score to get a personalized model, trained on your unique interests to recommend companies you might be interested in based on various factors including product and market.
Configuring Your Relevance Score
Access the Settings Page: Navigate to the settings page where you will find options to set up your relevance score.
Choose Relevant Company Lists: Designate the company lists to be considered when developing a score for relevant and non-relevant companies. You can add or remove companies from these lists or delete entire lists as needed. This flexibility enables you to refine your preferences over time.
Updates and Pre-defined Lists: The scoring system updates on an hourly basis, taking into account the preferences you have indicated in your lists. By default, Harmonic provides two pre-defined lists: a "liked companies" list for recommendations and a "hidden companies" list for non-recommendations. These lists can be removed or supplemented with additional lists according to your needs.
Integration with Polytomic: If you're a Polytomic user, you have the option to set up a reverse sync from your CRM to Harmonic. This designated list can help establish and inform the relevance score, ensuring that our models stay up-to-date based on the companies you find most interesting.
Viewing and Sorting by Relevance Score
Accessing the Score: Users can view the relevance score in a column designated “Relevance Score”.
Sorting: By default, the sorting is set to "Featured", which indicates generally relevant companies in the Harmonic ecosystem. Update the sort to Relevance to sort by your own custom relevance score.
Training Your Model
The companies in your designated lists are used to train a model tailored to your preferences. There is an enforced minimum of 30 companies. For the best results, we recommend including as many companies as possible in your lists and regularly reviewing your selections over time. The more companies you add, the better our model becomes at understanding your preferences and providing accurate recommendations.
Note: Lists of companies in your "Do not recommend" section will start impacting the score as soon as they are added and the model is updated. There is no minimum number of companies required for these lists to start affecting the score.
What if There Aren't Enough Qualified Companies?
If there are not many companies on your list, we can still generate a score, although it may have some gaps. The generated scores will reflect lower confidence levels due to fewer data points. On the other hand, companies with a large number of entries in their lists (at least 200 with recurring CRM sync) will have greater relevance.
Ongoing Development
Our team is continuously working to develop additional scores to improve the recommendations we make. If there are other factors you believe we should consider, please let us know!
If you need additional assistance or have further questions, don't hesitate to schedule a time to sync with our customer support team. We're here to help ensure that our models are as accurate and useful to you as possible.