Corybant was selected by the Colorado Department of Natural Resources Division of Wildlife (CDOW) to automate the 2005 Fall Turkey Survey and the Fall Ptarmigan Survey. After an extensive three-month competitive bid process, Corybant was then selected as the preferred vendor for automating the CDOW Annual Big Game Harvest Survey (2006-2010). For the Big Game Harvest Survey (Big Game), each year between October-February, approximately 160,000 hunters are contacted and at least 70,000-90,000 completed surveys are collected. The Big Game survey consists of approximately 2500 dynamically generated survey questionnaires (one for each hunt) that are personalized for each hunter and specific license based on the hunter profile. From 2005-2013 Corybant’s Automated Survey System was used to issue over 3-million personalized notifications, and collect over 400,000 completed surveys with 99.9% accuracy. In addition to automating Colorado Small Game Surveys from 2005-2013, from 2006-2010 Corybant automated Colorado annual Big Game Harvest Surveys and saved the State of Colorado more than $100k annually.
The timeline for automating the Big Game Survey
The CDOW Annual Big Game Harvest Survey (Big Game) project is an example of the efficiencies that are realized by using Corybant’s project management methodology and hosted database and application services. The Big Game project was awarded in August of 2006, the contract was signed on September 21, the first notifications were issued on October 4th, specifications were finalized on October 16, and the project was in full deployment by November 4, 2006. Over the next three months the system issued over 760,000 personalized notifications to 150,000 hunters and by February 14, 2007 the system had collected more than 73,000 completed surveys. During this timeframe, Corybant analyzed existing procedures and created analogous automated procedures and migration strategies to meet CDOW’s operational requirements and to transform legacy (human operated) survey questionnaires into analytical models that could be coded into web based questionnaires and IVR scripts. During the course of the initial year of the project, Corybant set up redundant encrypted databases that were used to: validate survey participation with the master contact lists (login, user ID, etc); log current contact lists (indicates time, date, method of contact and status of participant); verify survey responses (a number of verification points are implemented); and detect conflicting responses (hunt/did not hunt, harvest/did not harvest). The scope of the project included agreement with CODW on IVR and on-line survey questions for each form, creation of notification and call-list management strategies in collaboration with CDOW, implementation of an interactive web portal, a development portal for testing, and integration and training of a 30-agent call center, and implementation of secure database, IVR application and scripting, user studies, web interface, reporting and monitoring, and operations support utilities.
Background and importance of harvest surveys
The Colorado Division of Wildlife (CDOW) is now the Colorado Division of Parks and Wildlife (CPW), and is a division within the State’s Department of Natural Resources under the jurisdiction of the Colorado Wildlife Commission, that was created by the Colorado General Assembly in 1897. CDOW/CPW manages 960 wildlife species and over 230 wildlife areas. More than 1.1 million people hunt, fish, or trap annually in Colorado and infuse more than $1 billion dollars into Colorado’s economy each year. A CDOW economic model suggested that expenditures by hunters alone generated $500 million for Colorado’s economy on an annual basis. Seventy percent of the Division’s funding comes from the sale of hunting and fishing licenses. In 2005 the sale of hunting and fishing licenses resulted in more than $65 million in revenue for CDOW/CPW and of these license revenues more than 75% are from the sale of deer, elk and pronghorn licenses. As part of managing the deer, elk and pronghorn herds and allocating licenses, CDOW has conducted an annual big game harvest survey of Colorado hunting license holders since the 1940’s. In 2005, the DOW sold over 378,000 deer, elk and pronghorn licenses for over 2,300 different hunts.
Need for Automation
Prior to Corybant’s Automated Survey System all CDOW/CPW Small Game (e.g. Turkey, Ptarmigan) and Big Game harvest surveys were conducted manually via a 30-to-40-person call center who would contact each hunter by phone, record the responses on a paper questionnaire, manually enter the questionnaire data into a table, and submit the raw data entered to CDOW. Then CDOW would attempt to manually upload the raw data into their database (many rejections happened at this step) and once uploaded CDOW would perform error checking against the information recorded, reject any surveys with errors and send the rejected surveys back to the call center for re-processing. The manual processes for each step took several days and sometimes weeks, the error rate was extensive, and the cost of conducting the surveys was prohibitive.
Project Management Methodology
This project is an example of the efficiencies that are realized by using Corybant’s project management methodology. The project was authorized in August of 2006 and was in full deployment by November 2006. During this timeframe, Corybant analyzed existing procedures and created analogous automated procedures and migration strategies to meet CDOW’s operational requirements and to transform legacy (human operated) survey questionnaires into analytical models that could be coded into web-based questionnaires as well as IVR scripts.
The efficiency of the project management methodology is best exemplified by the two user studies that were conducted during September of 2006 where the results of the on-line user study was used by the subsequent study that was used with the IVR system and the overall results were rolled into the final product, deployed in October 2006.
The scope of the project included building an interactive web portal, a development portal for testing, preparation of training documents, and integration and training of a 30-agent call center, and implementation of secure database, IVR application and scripting, user studies, web interface, reporting and monitoring, and operations support utilities.
By October 2006 Corybant provided CDOW a multi-modal survey system for the Big Game Harvest Survey that managed secure access to the system for approximately 150,000 hunters and presented over 2500 dynamically generated questionnaires that are data-base-driven and personalized for each hunter. Corybant configured and implemented a secure database, branded survey website, branded IVR service, and a branded call center -- all seamlessly integrated for CDOW. The system continued to improve and evolve as new technologies and requirements were added.
Corybant’s Survey System configured for the State of Colorado DNR
A summary of the key and distinguishing features of Corybant’s Survey System as it was configured for Colorado included: a secure, branded, fully integrated website, a branded telephone number (COHUNTERSURVEY) and a dedicated toll-free number for phone surveys. The service included a seamlessly integrated automated telephone (IVR) system; and a Colorado-based help desk that was fully integrated with the online and IVR system and staffed 24 hours a day with up to 30 experienced operators familiar with the Colorado harvest survey and Colorado game management units. The system further includes a mechanism for capturing updates to hunter contact information, transmitting updates to CDOW, and utilizing that information in future notification attempts.
Corybant’s automated notification service is configured to implement policies that were specified by CDOW for the most effective notification strategy. The system evaluates each survey record individually and aggregates the survey notifications by hunter and household in order to maximize the success of each notification. It determines the appropriate contact method and timing based on CDOW policy for alternating notifications between email and phone based on the attempt and includes the ability to suspend notices for a particular survey, sample batch or other criteria.
Appropriate to each survey method used (online, IVR or operator) the system provides a single secure login for all survey participants that re-directs the participant to the appropriate survey based on their login information and is currently configured to present up to 4 surveys for each hunter in a random order based on stated CDOW specifications.
All hunter responses are logged by method, time, and date and operator-assisted calls are recorded for quality assurance and training. Corybant has configured the online survey with interactive maps that fully describe each GMU in accordance with landmark information that is provided by CDOW.
They survey system is integrated with the back end systems for dynamic generation of survey questions in accordance with such variables as the hunter, hunt code, and season.
The automated phones system includes a sophisticated IVR voice dialogs with intuitive voice prompts, text-to-speech, and real-time error checking based on CDOW policy for particular hunters, licenses, or question form. These dialogs include voice prompts that effectively “walk-the-hunter” to the correct GMU or connects them with a live operator if they press “zero” or if the system detects they are having difficulty. In its current configuration the IVR system is able to process over 600 hunter notifications per hour with instant access to up to 30 trained help desk operators 24 hours a day.
Of particular note is Corybant’s work on the database and reporting aspect of the system. Since 2006, Corybant has implemented processes in line with requirements specified by CDOW, for accepting and reporting hunter sample lists by electronic transfer (including capturing and reporting updates to hunter phone numbers and email addresses). Corybant has implemented specific CDOW requirements for additional information such as identifying the transaction history (including survey method or methods used to complete the survey), identifying the notification history (including the timing, alternating contact methods and greetings, aggregations, etc. for each notification) and correlating the contact attempts with the actual survey to support CDOW in identifying and eliminating biased information.
Corybant’s survey system utilizes a secure distributed network that is available 24-hours a day and includes fully redundant automated voice systems with trained help desk assistance 24-hours per day.