Adoption Matching and Process Improvement Final Report
Adoption Matching and Process Improvement Final Report
The client, Atlanta Humane Society takes care of animals that are rescued and then adopts them back to people.
The project focused on improving the routine care taking processes of washing and feeding of animals, the adoption process itself, and the notion that we could make recommendations to adopters to lower the return rate.
This is relevant to the project because a significant amount of time is required to take care of animals, because in the adoption process the wait times are high, and because the pet return rate is quite high.
The three problems we worked on are as follows:
1.) There was no time study done on the system carrying out of routine care, which makes it difficult to monitor the systems performance and suggest opportunities for improvement.
2.) The wait times in the adoption center were quite high.
3.) The pet return rate is quite high, and there is no system in place to inform adopters of what pets could be a good match for them.
Opportunities we identified:
1.) After conducting a time study, we found that there was a lot of idle time in the feeding process, and that joint stall cleaning was slower than independent stall cleaning which were both opportunities to make the process better.
2.) A lot of paperwork was required to be filled on site by adoption staff and adopters that could be done online beforehand and save time.
3.) Create a system based on past data that could help inform adopters of what sort of animal might be a good fit for them.
1.) Process improvement - For recommending independent washing, simply present the data that the process would benefit in time saved by advocating for this policy.
For the feeding process the methodology for developing the design of the batch feeding system was by creating time saving estimates to the AHS staff based on the results of the time study, suggesting this as a solution, working with AHS staff to formulate how it might work, and then putting our plans to action in the testing phase. During our testing which was limited by the COVID-19 to 1 trial, we found a time decrease for the employee who was a part of the trial of 50%.
2.) Adoption Process -
The methodology we used to develop the design was to create a simulation of the adoption process on the busiest days, Sunday and Saturday. The data for this simulation was generated through a time study we conducted. Once we obtained corresponding performance in our simulation model as might be expected in the real system. We tested the difference in the system if it were the case that certain paperworks were not required to be filled on site and the result was estimates of total time in system reductions from 4.1% to 20%, resulting in up to 236 hours of time savings for adoption counselors per annum.
3.) Recommendation -
The methodology we used to develop the design was a combination of what is called the affinity propagation clustering method and a logistic regression done on the data of past returns. The design itself was created by working with AHS and finding a marketable and implementable way of using the results of our analysis to help adopters find better matches. We decided on a survey to categorize adopters and match them to the categories the data analysis on the pets returned. We were planning to test and validate by piloting our recommendation model, and seeing how it performed briefly before the final report but to due complications related to COVID-19 this was not possible
1.) Meal batching process document containing system formulated with AHS staff based on time study results and tested to lower feeding time.
2.) Recommendation for independent stall washing based on time study results which showed that this method for washing would lower washing time.
1.) Simulation model of Saturday and Sunday adoption process with and without "Fast Lane", what we call taking paperwork online.
2.) Demo of the main application form being transferred to an online setting, facilitating ease for them to move further forms to this setting as desired.
These deliverables facilitate lower time in system for adopters and save adoption counselors valuable time.
1.) 10 survey questions that provide potential adopters a recommended music genre or category in which they fall to base adoption decisions.
2.) 3 survey questions that provide potential adopters a readiness level which is another category which evaluates their preparedness to take care of a pet belonging to one of three categories of pet ownership difficulty.
3.) A demo of the pet matching model.
4.) An animal music genre assignment tool, and the Excel code to support this.
These deliverables help provide an adopter a framework for understanding what kind of animal may be good for him or her given his behavioral tendency and experience as a dog owner.
Value and Impact
1.) Routine Process Improvement and Time Study:
120 hours of feeding time saved per annum.
70 hours washing time saved per annum.
Time study shows that each percentage decrease in return rate from 3.) results in 65 hours of staff time saved per annum.
2.) Adoption Process:
Total time in system for weekend adoption center guests reduced by 4% - 20%.
Max time in system for weekend adoption center guests reduced by up to 12.8%.
Counselor utilization in weekend adoption center reduced by 38%, perhaps giving them chance to contribute to AHS in other ways while still getting done all of their responsibilities.
Counselor time saved ranging from 44 - 236 hours.
3.) Recommendation Model:
Up to 86% reduction in mismatches of animals and owners
Potential decrease in return rate for dogs from 11.7% to 5.4%