Integration and Data Synchronization
- Our connector for commercetools is able to collect data and keep it in sync with our backends automatically. However, for any reason, should you decide to provide data via external feeds, we request the data feed to be provided in Klevu’s format (see /hc/articles/39396957275035). In this case, feeds must be accessible via a public URL. If needed, the Klevu team will be happy to provide a list of IPs to whitelist for access. If you want us to download your feed at a specific time of the day, we request you to inform us about the same.
- Regardless of the platform you use, we expect you will try Klevu on your staging / preprod environments prior to installing it on your production environment.
- We also expect you to have gone through our generic checklist and taken the relevant actions before going live (see /hc/articles/39397050457499).
- For the front end integration, you have three options. You can use our SDK (recommended), ready to use Klevu theme or Search API (see https://developers.klevu.com/). Should you wish to write custom search queries for specific use cases, please share with us your use cases and our search team will be happy to help.
Analytics
- If you are integrating search, catnav and/or recommendations using our API (as opposed to using our SDK or our Klevu theme JSv2), you must implement all the analytics reported in our analytics API documentation. Please see https://developers.klevu.com/api#analytics.
- On commercetools, you must implement the following endpoint (see https://www.klevu.com/developer-api/api/reporting.html#checked-out-products). If you are using our Headless SDK (see https://headless.developers.klevu.com/sdk), please implement a checkout call using the respective SDK function.
Self-Learning
- Our Machine Learning models for self-learning are retrained every other day. The models, by default, use the data of the last 30 days. Should you wish to make changes to any of these parameters, please let us know.
- We recommend that you provide us with historical data. At the minimum we need data of the products purchased in the last 2 to 3 months. We expect the data to be made available as per the guidelines provided at /hc/articles/39397012198299.
- If you intend to perform A/B testing of Klevu with other systems, it is very important that we receive your historical analytics data including the products bought and visited by individual customers as well as the search queries fired by them. Please, also share the KPIs that you’re planning to monitor when performing A/B testing.
- If you are using google analytics,
We recommend you follow the guide below to enable submission of Klevu analytics data to your google analytics account (see /hc/articles/39397051012891). Please also give us access to your Google analytics account (use contact@klevu.com as the account). We need this to compare website performance before and after Klevu and take necessary actions if need be. - If you are using any alternative to Google Analytics, we request a short demo of your system and possibly a read-only access.
Relevancy and Performance
- Customer queries are usually formed of product nouns (i.e. chair, sofa) and attributes (e.g. color, size, shape, material). For Personalisation and Recommendations as well, attributes play a very important role in personalizing customer experience. It is therefore very important that in your catalog you have product attributes included.
Customizations and Compliances
- If you are using our Klevu theme for the front-end integration and if you have requested the Klevu team to build / customize UIs, we expect to receive your requirements with user acceptance criteria. Our UI team will not commence any work until the respective wireframes / images accompanied with functional specifications have been signed off by the team.
- Where you have requested specific UIs to be implemented, Klevu will not be responsible for any patent infringements related to functioning of the UI interfaces.
- Should you need your UIs to be compliant with any specific requirements, we expect you to share them with us with all the necessary details and documentation as part of the functional requirements.
- All UI/UX or any custom work will be chargeable on hourly basis
Multi-tenancy and Dedicated Hardware
- Unless otherwise agreed, catalog indexes and analytics data are hosted/processed on multi-tenant systems where other customers’ data may be stored as well. Should you require dedicated hardware, please let us know. Dedicated hardware is usually available as part of an enterprise plan and at an additional cost.
- In order to respect data privacy policy for all our customers (including yours), it would not be possible for us to give audit rights to any of our multi-tenant systems. However, incase of dedicated hardware, we would be able to provide access to only the search systems.
Dealing with Sensitive Data
- Please note that any content provided in the feeds is indexed as it is at the Klevu’s end without any filtering applied to the content. This information is searchable and therefore visible in the search results when queried for. We may add additional information to the catalog to enhance your catalog, however, for the content submitted by you, it is, therefore, very important that you review and decide what content is sent over to us via feeds. Klevu can not be responsible for storing sensitive information that was provided to the index by the customer.
- Except for the details already mentioned on our GDPR page (see https://www.klevu.com/gdpr/), Klevu does not collect any customer’s personal information. It is important that you too are not sending any content sensitive to your business and/or any personal information of your business and/or customer to us via API calls or feeds.