Skip to main content

Frequently Asked Questions

What is Cherry Recommends?

Cherry Recommends is an AI-powered recommendation engine, which is a system that recommends items, services, and information to users based on advanced data analysis and machine learning. Recommendations are chosen using a range of criteria, including the user's history, attributes, and comparable users' behavior.

How does Cherry's machine learning work?

With built-in machine learning, Cherry continuously analyses interactions with recommendations over time and automatically improves it's algorithms to enhance your outcomes. It utilises a range of clustering, segmentation, reinforcement learning and Bayesian techniques.

Does Cherry change my revenue instantly?

Cherry will start personalizing the offers given to each customer as soon as deployed. You should start to see statistically significant results within 2 weeks of correctly deploying Cherry with sufficient data.

Your revenue should consistently increase for 3 to 9 months after until the AI achieves an optimal result. Changes in customer behaviour (e.g. post-Covid buying behaviour) may quicken or delay the performance improvements.

Is Cherry able to provide recommendations for anonymous customers?

Yes, Cherry can still provide recommendations for anonymous customers but they won't be personalised. Cherry assumes that all anonymous customers exhibit similar behavior.

How many customers do I need before I can use Cherry Recommends?

You should have at least 500 customers per week going through the lifecycle stage you want to personalise. The AI will work down to 100 customers per week but performance may drop off. For example, lets assume you want to increase your free-to-paid conversion rate. You will want at least 500 customers per week (or ~2000 per month) starting a trial.

How fast is Cherry?

99% of recommendations are produced in less than 1.4 seconds.

Can someone just keep clicking and get different Recommendations?

Cherry caches (keeps) the same recommendation for a customer for a set period of time. You can set the cache length as 1 hour, 1 day, 7 days or 14 days.

Can Cherry handle multiple users within a single customer company?

Yes. We'll update the docs soon on how to set this up. For now, just ask us and we'll set it up.

Is Cherry customer journey based OR is it more campaign based?

Cherry was developed to retain customers throughout their lifetime rather than be campaign-based. However, if running a particular campaign within a specific period, recommenders can be set up as campaigns within a defined period with unique offer parameters.

How does the discount work for checkout? Do discount codes carry through?

Discount codes presented can be configured to carry through to the checkout page.

Should discount codes be in absolute or percentage terms?

Cherry Recommends can handle both. Just make sure you describe it in the properties of the item.

What do I need to make Cherry work?

Setting up Cherry is a five-step process:

  1. Connect customer data source, e.g. Segment

  2. Add in items that you want to recommend

  3. Set up Recommender

  4. Deploy Recommender

  5. Building your recommendation UI

How do I feed data to Cherry?

Cherry integrates with other systems. There are two types of connection:

  • A connection to a third party web service like Segment or Hubspot

  • A custom connection, e.g. your own web service.

Is my data secure with Cherry?

With Cherry, your data remains encrypted and secured at rest and in transmission. Cherry has 256 bit AES encryption, OAuth compliant secure user logins and customers are analyzed anonymously without personally identifiable information (PII).

What other benefits does Cherry provide?

Upon request, Cherry is able to provide optics over your customer segments and identify who are your 'high lifetime value' customers based on various attributes (not just current spend). These customers can be used for look-a-like audience marketing efforts so you can attain even more high-quality customers.