I've researched and written a lot recently on search and discovery solutions. That research led to a conversation with Pierre Bergeron of INO (Institut National D'Optique) about their deployment of Coveo's Search & Relevance technology. As Process & Compliance Manager, he has an in-depth knowledge of the productivity gains INO got from implementing the Coveo suite of unified search technologies.
Coveo's Search and Relevance technology harnesses data from any source and recommends knowledge and subject experts relevant to the context of the user. This solution provides access to systems both inside and outside the corporate firewall, and to data that is both structured and unstructured.
Coveo's products are used by over 700 customers in a wide variety of industries, including Lockheed Martin, Harris Corporation, Haley & Aldrich, Sony Ericsson and the U.S. Navy.
Founded in 2005, the company's North American operations are located in Quebec, Canada (R&D) and San Mateo, CA. Coveo recently opened its European headquarters in the Netherlands. The total employee count is about 150.
INO
The company is a nonprofit organization, based in Quebec, Canada. INO was founded over 25 years ago as a research and development center for technology and product development in the fields of optics and photonics.
Though INO's primary focus is R&D, the company also offers prototyping and short-run production of custom products. Through technology transfer, other companies undertake further development and commercialization.
Scope of INO's deployment of Coveo
Pierre said, "To understand our approach, you need to know how we defined the scope of our life cycle process. We adapted the Technology Readiness Levels definition from NASA. It identifies nine steps; TRL 9 is technology that has been proven through successful operations. TRL 6 is significant to us; it affirms that a product has been successfully demonstrated. Our policy is that we will not cross to the next step toward commercialization, TRL 7, without customer engagement. "
Pierre and I then explored how Coveo supports INO's broad range of activity. INO uses PTC Windchill and Pro/E products for product development and Microsoft products for ERP and CRM. It is accessing these and other related systems that brought about almost complete employee use of Coveo. In addition to ERP/CRM systems, the application of Coveo accesses many different types of documents, such as PDF, PowerPoint and Excel, and file systems. SharePoint 2013 is an example. Image searching is also possible.
Pierre said that their Coveo implementation provides access to 99% of the "searchable" content.
Pierre pointed out several ways INO personnel use Coveo's search power, "One is a basic keyword inquiry, usually for a specific need, phrases, or more complex search syntax with Boolean operators such as AND, OR, parentheses, etc. Other ways include what I call Knowledge Management, more suitable in the early stages of the TRL scheme, and Business Intelligence."
In the Knowledge Management area, INO has restrictive security procedures to guard their Intellectual Property. A typical search might answer questions like what technologies have been used for which project. Not all inquirers can get requested information, although with Coveo, they can see that the information exists and who owns the document, so they will be able to ask for permission to access it.
Pierre continued, "To extend our knowledge Management, we are planning to acquire/deploy the Coveo Text Analytics module, which will bring additional significant value. As our personnel have experienced, an inquiry may bring up a huge list of possible relevant sources. The facets and filtering capabilities of Coveo help us to refine our searches."
Measuring ROI on Semantic Search at INO
ROI is important to the management of INO. To gauge possible productivity gains, INO did a Proof of Concept on a very limited set of data. Pierre said, "We could recognize some hard dollar savings as well as some softer benefits. Our estimate of hours to be saved was a very conservative figure, yet because of the use of the Coveo system, the early productivity gains were substantial." In the end, the high use rate of 99% of all employees makes it easier to add up the hours saved.
INO set three major goals for the implementation:
- Rapid response time of under one second on any searchable data
- Stop sorting and trying to manage email files
- Support the Knowledge Management needs of the research people
In the Knowledge Management area, the justification numbers get soft. Pierre and his team knew that there is a cost to avoid re-doing what already exists, but they could only describe those situations anecdotally.
The hard dollar savings showed up in a total productivity gain of 5%, so the INO breakeven point was less than one year. Also, there were some intangible benefits such as avoiding the frustrations of extended searching and not finding results, or bothering and interrupting colleagues.
To ensure that these benefits continue, INO immediately trains new employees on using the system, usually in only half an hour. Pierre reports that having access to the system helps new employees get up to speed faster, and easily learn who does what in the organization.
Uncovering the value locked in email files
INO's initial implementation focused on the rich contents of the extensive email files. These files were built up over many years and contained valuable information for product developers. The implementation team faithfully adhered to the Pareto, 20-80 rule to drive to their ROI goals. The cross linking to ERP and CRM came next, again expanding the use of the system to the Supply Chain.
On the cost side of the ROI equation, INO avoided unnecessary implementation costs by staying standard with the Coveo software.
Pierre recapped by saying, "Efficient search is the number one need for every employee in Knowledge Management and day-to-day work since they are looking for information 5 to 30 times a day. Use it, you'll love it!"
My thanks to Pierre for sharing INO's experience implementing an efficient search solution.