Constructor.io Search vs. Coveo Relevance Cloud

Constructor.io Search takes all of the guesswork out of maintaining a search engine. As merchandisers or product teams, we have educated guesses at how search relevance should work but the customer is always king. We can't always predict the ways in which consumers will search or what their intent is. That's why the behavioral-driven approach that Constructor.io employs works so well. It means that merchandisers can focus on their sales and promotional responsibilities, instead of wasting time and bandwidth on base-level relevance questions.

Incentivized

Coveo Relevance Cloud is a great solution to implement into Salesforce to provide Knowledge-Centered Support, Enhancements to a Customer Community, to provide sales aids, or to complement your customized app in Salesforce.

Incentivized

• Optimizes for business KPIs, not string matches, so merchandisers can focus on strategy instead of maintaining a dictionary.
• Gives merchandisers control to curate the right customer experience per their expert understanding.
• Customer/technical support during and after implementation.

Incentivized

• Coveo is fast, search results come up quick (though it's not always great).
• Not much complexity to run.
• Coveo is implemented within our portal and doesn't require extra steps to use it.

Incentivized

• Visibility of the personalization algorithm.
• Exposing detailed analytics within the customer-facing dashboard.
• "Top searches" available via API.

Incentivized

• It would be great if Coveo 6 allowed you to rebuild indexes from a certain subtree instead of needing to rebuild the entire tree to see changes. This functionality was added in Coveo 7 and is very useful.
• In Coveo 6, integration with Sitecore is more difficult than one would expect. This integration is much improved in Coveo 7.
• I have seen cases where an exception thrown when crawling a specific document will cause the indexing to stop completely. I believe this only happens in implementations using custom faceting but it could be handled more efficiently if the trouble document was skipped and the indexing could continue.
• Relevancy ranking editor is good but not as powerful as GSA. GSA offers a self-learning scorer which automatically analyzes user behavior and the specific links that users click on for specific queries to fine tune relevance and scoring.
• We've ran into issues on multiple clients with Sitecore items being indexed multiple times in Sitecore 7 and Coveo 7. The fix Coveo suggested was to upgrade our Sitecore version and Coveo but unfortunately this didn't resolve our issue. After months of testing we were finally able to resolve this by implementing our own CoveoItemCrawler to get around the issue (based on https://developers.coveo.com/display/public/SC201404/Items+in+the+Same+Language+Gets+Indexed+Multiple+Times;jsessionid=3C1A2AE33540E0A0B8BB52BA3A64AF70).
• Integration with RabbitMQ in Coveo 7 seems error prone. We often see the error "The AMQP operation was interrupted" and on occasion, need to restart the Coveo service to get this operating again. In some extreme cases, we have also had to restart the server because of issues when attempting to restart the Coveo service.

Incentivized

This question is not applicable to me

Incentivized

• Quick to find things in a massive database when needed.
• Results need to be more concise - sometimes we spend more time looking for the right file than if we were to just search amongst our own networks instead.
• Coveo is not always the most useful but does its job when general information is needed.

Incentivized