Webinar presenter Dr. Josh Fisher answered a number of your questions after his presentation, Data-Driven Decision Making for Animal Welfare Organizations Part 2. Here are just a few of his responses.
Audience Question: Josh, can you go back a little bit? Can you explain again the difference between predictive and causal? Are they the same?
Josh Fisher: Sure. So, with predictive, we’re looking at based on this information set that we have, you know, what is likely to happen next. The causal we’re actually looking at trying to establish that cause and effect. So, they are very similar, but they are distinct types of analysis. The predictive we’re using existing data to try to establish what our inventory is going to look like. I’m trying to think of other good examples, and I apologize I am drawing a little bit of a blank. But where we’re using existing data. We’re using the things that are happening. That have already happened or our historical data to predict what’s next. Cause and effect, candidly, in an animal welfare setting, we’re usually looking at the effect and working backward to the cause, right? So, if we’re looking at our illness, we have an outbreak of distemper within our shelter and we’re looking through trying to figure out and trace back to what that root cause of the distemper is. We can run reports to see, this patient broke with distemper. Let’s look back and try to figure out from this animal what different kennels it was in, and can we tie it back to another animal? Can we tie it back to a group of animals that came in together that were all sick? Can we tie it back to an animal that was vaccinated later upon intake, etc.? So, in an animal welfare state situation, the predictive analysis is usually going to be using our historic data. So, in May of 2019, 2020, and 2021, we had animal inventories of this. So, for May of 2022, we need to make sure we have X number of vaccinations, X amount of food, etc. on hand, right. Then, the cause-and-effect piece in animal welfare, generally speaking, what we’re looking at for that is going to be the effect of working backward to the cause. So, it makes sense.
Audience Question: I think so, and then so, there’s another term out there called correlation. How does that fit into all of this?
Josh Fisher: So, establishing the so correlation. Correlation is not always causation, right? So, especially in animal welfare, that correlation is going to tie really closely to what we’re talking about with cause and effect. It’s looking at how the two things are tied together so that, again, we can use that exact same example for our distemper outbreak within the facility. You know, is there a correlation between the animals coming in from XYZ porting case and there being an uptick in distemper in the facility. Were those animals vaccinated upon entry, were they vaccinated prior to entry? So, correlation is going to be more in line with our cause-and-effect approach, than the predictive piece.
Audience Question: Are you aware of any shelter care systems that have sophisticated reporting already built into the software?
Josh Fisher: So, yes and no. All of the shelter care systems have a pretty substantial, customizable aspect. So, Chameleon, Pet Point, Shelter buddy, etc. You can have those different preloaded reports entered into the system and have those dump out information to you. They can run very complex calculations, right? It’s all a matter of how they’re coded. And, again, the data that is being put in as to what the accuracy. So, we come back to that, you know, garbage in, garbage out concept. So, what we’re looking at with the complexity being built-in is… A, what kind of data and how much data are you already capturing in your system? And then B, can you work with the software providers to get those reports run and ready to use for you? And a lot of that is going to come down to… A, your preferences. So, you want the same report and the same format every single day? For example, in Chameleon, we have set up, with using that Chameleon system, plus that their kind of postmaster function that we get inventory reports and then euthanasia from the day prior e-mailed to the management team here every single morning at 8 AM. So, those reports so that if something absolutely crazy happens and, you know, I’m seeing 18 euthanasia is from the day prior, A, my hackles are going to go up because I’m super confused as to why that many animals were euthanized but B, it gives the information as to why they were euthanized. So, if something looks really off. Like I have you know, a report that pops up and says that it no, we had an animal euthanized because of space or something and I’m looking at the facility. I’m, like, okay, well, we have plenty of space, so that shouldn’t have been a thing, right? It helps you to detect issues early, and that is a super simple report. But the complex analysis piece, you know, again, it’s going to depend on how much time you have, and the company and you work together to dive into creating those reports. I don’t know how many of you have seen the new Shelter Animals Count Reports. But that’s a really good example of a report that requires some complex coding and calculation in it to pull that data and display it. But the new reports have even more information than the old ones. So, if you’re spending the time and really going to dig into that information, there is the capability. It’s a matter of doing that backend coding to get it done.
Audience Question: Josh, we’re a small rural agency, and just barely have enough people to go out and pick up the dogs. Do we actually need to hire somebody specifically to do this analysis? Do we have to have certain skill sets, degrees, to basically do what you’re describing or is it as simple as learning the software systems?
Josh Fisher: It can be as simple as learning the software systems. So, the thing that is most beneficial to your organization is first identifying what your priorities are, what do you want to see, and what information is going to help you with your decision-making. So, if that information is your highest intake zip codes. If that information is identifying what your most at-risk breed is. If that information is drilling down into your length of stay. With our cats, for length of stay, is it the kittens? Is it the adults? Is it black cats? You can drill down as specifically as that. And you can do that without having to have a whole lot of additional skills and data manipulation or anything like that. Like I said, I’m an Excel person. That’s kind of how my brain works, but if you wanted to know, for example, what’s the average length of stay for cats broken down by color. You can talk to your software provider, and you can get them to assist you in writing that report so that all you’ve got to do is run it. You don’t have to be the one that does any of that coding. You just have to have the question in mind, and then work with them to get that information in an easy-to-read format. That’s part of the advantage of paying for these shelter softwares.
Audience Question: Josh, are there standing reports that you pull or write, about always saying it every month, or on a set time schedule? So, if so, what are those reports, especially the ones that you feel might be beneficial, especially if we were starting from ground zero? What are those standing reports that maybe we should be thinking about?
Josh Fisher: Sure, we do run and report on our shelter animal count data every month, and that’s definitely kind of a foundations level report. We also look at on a monthly basis, at our length of stay report so that way, we can identify for some reason, did dogs tend to stay longer this month than they did last month? And if so, can we look into why that happened. We do here run our euthanasia report every month, and we run the details of that, so we can figure out, did we have a lot of hit by cars this month? Did we have a lot of cancer? Did we have a lot of owner-requested euthanasia? You know, what was our big reason for loss of life, essentially, within the organization? And then from a Field operation standpoint, we do those high intakes zip codes every month. We look at our return to owner rate, that’s a big one. So, we run return to owner in the field that we have that coded in the system. One of the big reasons that we look at is, you know, good, bad, or otherwise, for officers, you can see them slip, if you will, into the easiest thing for them. Is to just pick up the animal and bring it in. You know, it’s easier to do that than it is to go knock on doors and ask questions. So, if we see our return to owner rate start to drop, then our field operations Manager and our field supervisors will have some conversations with the officers about, “Hey, guys, you know, remember, this is a priority for us. We want to make sure we’re returning them into homes rather than bringing them to the shelter as much as possible.” Let’s see what else, what other reports do we run. We always run our adoption report, we’d like to see what animals, what types of animals are moving. Our veterinarian is very interested just on a personal level as to cats and kittens, and what colors tend to be moving faster. And it’s always interesting to see. Generally speaking, Siamese flies out the door, if it even kind of sort of looks like a Siamese. And so anecdotally, again, you know, we in the industry can kind of predict, but it really depends for your organization, what your current priorities are. So, our monthly reports, and what reports we run on a frequent basis, are going to depend on what strategic initiatives we’re focusing on. And what our strategic plan looks like for that year.
Audience Question: So, it sounds like what you’re saying is, go back to your organizational objectives or the key strategic priorities that that agency has? And then tie your data analysis, or your data activities, or your data gathering activities to those organizational goals, objectives, and strategies?
Josh Fisher: Absolutely.
Audience Question: Josh, do you have any standardized reports for other activities like marketing and social media?
Josh Fisher: We do. So, I will be 100% honest as we have just recently started using our shelter management software for tracking our development team, so fundraising. So, we have started developing some reports for donors and donor tracking. Do we have repeat donors, if so, what frequency and what amounts? We also have some standardized tracking with our social media. We are getting to a point where we’re able to coordinate our foster parents in particular because they tend to share and repost and flagging them in our sheltering system as active participants in social media. So, that when we do push out an urgent plea, for example, on social media, we have a quick list of people who are actively sharing our information already. And we can send an e-mail blast to and say, “Hey, guys, we’ve just put out this urgent plea. Can you please share it?” We send to our volunteer base, but looking at our active followers and flagging them in the system and being able to kind of keep that list separate has helped us to get some additional shares and additional reach.
