Finetuning Timeouts

Overview

When integrating SEON into your workflows, you have to consider how long you can wait for a response from our system. While you can control some elements of this timeframe, others are down to external factors.

The following sections offer an overview of what timeouts are, how they work within SEON, and how they can affect the results you see. Make sure to monitor your results in the first few days and make any changes you feel are necessary to get the best results. You can also reach out to our team at any time, with any questions that you may have.

What are timeouts?

In an API-based service, timeouts define the total time available for a request to be processed and answered. Total time means the time it takes for your request to travel to our servers, to be processed and answered, and the time it takes for the response to travel back to your server.

In SEON, response time includes three major components:

• Network latency: The delay in communication over a network with our APIs.
• Data Enrichment: The time to get all values for Data Enrichment. As SEON does not control third-party servers and data sources, response times may vary.
• Scoring Engine process: The time to process all rules and custom configurations in your account to create the full API response.

These three factors make up the total response time, otherwise known as the Integration timeout.

Note: All timeouts in SEON are measured in milliseconds (ms).

Setting up timeouts for SEON

When you run checks in SEON, our REST API requests data from the 50+ social sites we currently support. Each of these services has an optimal timeout range which ensures that we return information to you.

As a result, the timeouts you set up for your SEON integration will affect the accuracy of the information you receive. Because every business and use case has unique data needs and varying risk appetites, we've made changing timeouts as easy as possible.

You can configure the integration timeout in your back-side API implementation. This will define how long your system waits for a response from SEON.

In addition, you can also configure the Data Enrichment timeouts (for the Email and Phone APIs) on the SEON settings page, or through your API configuration.

Note: The API configuration will overwrite the timeframe set on the SEON settings page.

Keep in mind that the Email API / Phone API timeout setting only affects one of three components. As a result, it will not guarantee a response in a specific timeframe.

When configuring your it’s best to add 500-1000 ms to the Data Enrichment timeout you’ve defined in SEON. This will allow your system to account for any network latency before dropping the request.

We highly recommend that all integrations be prepared to cope with client-side timeout errors. This is to ensure that transactions with no SEON response are not blocked in the event of unexpected errors or degraded API performance.

To ensure a stable user experience, we recommend that you accept transactions if your system does not receive a response within the timeout limit set. Transactions should only be blocked if the response received from SEON includes a state or score that indicates the transaction should be declined. For further information, please refer to our whitepaper on dealing with high scores and states.

Finetuning results with timeouts

Finetuning your timeout score is an important part of getting the best results from SEON. Increasing timeouts will affect the accuracy of Data Enrichment. Keep in mind that diminishing return is at play. At lower values, a 100 ms increase can improve the hit rate by 10–20%. However, at higher levels, a 500 ms increase may only provide one extra checked service.

You can even set different timeouts for various use cases. For example, if you integrate SEON at multiple touchpoints you can run checks at registration, login, and at purchase. 

Increasing the timeout at high-risk interactions such as registration and purchase can offer you (and your customers) the best security. However, keeping it low at less risky touchpoints, such as login, can ensure a frictionless experience.

The tables below will give you an overview of what to expect from different Data Enrichment Timeout settings with an accuracy of about 95%:

All sites and services listed only under the first timeout they will respond within. A timeout of 3000 ms will include responses from all services listed in the column and to the left.

The lowest supported timeout setting is 1700 ms, and only the most risk-sensitive teams will need a timeout above 3000 ms. For most use cases, setting the timeout threshold to 2300 ms is a good place to start.

Tip: For a step-by-step guide on How to Adjust Data Enrichment Timeouts, visit our Knowledge Base.