Transforming Loss Prevention Analytics & Shrinkage In Retail With Exception Reports

Exception report Speaker

Transforming Loss Prevention Analytics & Shrinkage In Retail With Exception Reports

At this year’s NRF Protect show I had the opportunity to share the stage with DSW’s Jordan Rivchun, Director of Loss Prevention.  DSW has been using Profitect since 2014, and successfully made the transition from traditional exception reports for POS to prescriptive analytics using Profitect’s Sales & EBR module just recently this May.  If you were unable to attend the conference, or were at NRF but missed this session I highly recommend you scroll down and watch the video of the presentation.  In this video, you will learn how DSW is using data and prescriptive analytics to reduce shrinkage in retail and grow as an organization.  Find out how Jordan’s loss prevention department has become integrated with, and invaluable to, inventory control, allocators and operations.  Profitect has been the tool that has allowed DSW to redefine how they look at shrinkage in retail by using prescriptive analytics to create a self-sustaining system of identifying profit opportunities and acting on them appropriately to drive success across the enterprise. Click here if you are interested in learning how to accomplish this at your organization.

If you would prefer to read the conversation, a full transcript can be found below.

1. Guy Introducing Jordan.
2. Jordan’s Opening Statements.
3. How Prescriptive Analytics Helps Small Teams.
4. Traditional Exception Reports VS Profitect’s Exception Reports.
5. Moving To Pull Methodology.
7. Data Management With Profitect.
7. Stopping Shrinkage In Retail With Exception Reports.
8. Prescribing and Managing Actions With Profitect.
9. Ease of Implementation.
10. Profitect’s Inter-Organizational Impact.
11. Predicting Shrinkage In Retail With Prescriptive Analytics.
12. Machine Learning & User Feedback For Improvement.
13. Examples of Success With Profitect’s Exception Reports.
14. Questions and Answers.

Guy Yehiav Introducing Jordan Rivchun.

Jordan is the head of LP at DSW  has been a long time customers of Profitect, using our inventory solution for a while, and just recently about two months ago they replace their legacy EBR solution with Profitect’s EBR solution and so I’m very happy with all the stories behind it and how we made the change so quickly and all the great insights.  With that, I would like to welcome Jordan to the stage.

Jordan’s Opening.

Thank you. All right do we have any shoe lovers in the house? Show of hands just a few okay. So we’ve got 23 million rewards members at DSW so we’ve got a few of you guys in the house which is fantastic like Guy said my name is Jourdan Rivchun and I’m the director of loss prevention for DSW. For those of you that don’t know DSW we’re about 508 stores across 43 states, we also operate the footwear business for Steinmart, Frugal Fannie’s, and Gordon’s as well as our lease business and we’re co-owners in Town Shoe  Co., Shoe Warehouse and DSW in Canada. So somewhat of a global footprint and as some retailers are continuing to kind of dwindle we continue to invest in growth and that’s really how Profitect has helped us use data to grow as an organization.

How Prescriptive Exception Reports Help Small Teams.

You know I think I would be remiss if I didn’t mention the size of our team. I’ve got ten people focused on the field for our organization and as I just looked at the most recent NRF statistics that’s 5.5 people for one billion dollars in sales, which is a little bit lean and Profitect has really helped us leverage data and that technology to essentially act as a force multiplier. So as Guy mentioned we’ve had the Profitect solution since 2014, really focused on the inventory module, planning allocation damages, return to vendors, what-have-you and I’d like to use the term we’ve hijacked the data. You know historically inventory is not necessarily the role of loss prevention, but we saw great insights as we implemented the solution and really leverage that from our inventory control partners and since ‘14 we’ve really taken off with the data. Not just within the loss prevention function but really with the whole entire organization.

As recent as May of 2017 we retired our legacy exception reporting tool which many of guys probably have in your environments and implemented as Profitect calls it the sales (and exception reports) module integrating both point-of-sale T log data as well as our e-comm data from DSW.com and any in-store transaction.

Traditional Exception Reports VS Prescriptive Analytics.

So when Guy actually approached me to speak today, you know I wanted to really help educate my colleagues and peers with you guys as to how I saw the difference between a traditional exception reporting solution and a prescriptive analytics tool in something like Profitect. You know traditionally exception reporting is going to provide a lot of false positives, you’re going to have people spending time, generally, four hours digging four hours taking action, and really what prescriptive analytics has done is it’s put the data into action and remove those false positives. And I’ll get into some of the change management here later in the in the conversation but again we’re drowning in reports and I’m sure you guys are drowning in reports in your environments as well and reports have great data but they don’t inspire action. They don’t give you consistent approach in response to the problems that you guys are having and that we’re having at DSW in our organization.

Again the exceptions are not descriptive, they’re not actionable and those false positives quite frankly waste time with five regional LP managers spanning 508 stores they don’t have the time or the resources to worry about digging. So we’ve got to give them the action immediately at their fingertips and that’s really what the solution has done.

Moving To “Pull” Methodology & Accountability With Profitect.

Again growing up in loss prevention it’s really about that pull type of reporting so you’re going to come in on a Monday you’re going to pull the reports that you want to look at you’re going to find those exceptions. Some of those exceptions you might have to email out to a store and get a response back we’ve actually eliminated that with the solution. So the push type of reporting has really increased the span of control for our regional loss prevention managers, investigators, and as well as put kind of an accountability, Which is one of our values as an organization right into motion. So when you think about accountability, how do you know that report or that action is actually getting to somebody?

Well, this solution has enabled us and me as a leader in my department to have visibility down to the user level at a button’s click. I can see what actions are taking place I can see resolution notes opened, closed, in progress and at a moment in time I can understand what the workload is and the follow-up that’s needed from my team.

And again from a change management perspective that’s vastly different from your legacy exception reporting solution where those users are just digging they’re stuck in this pool of data trying to find something. So again taking our users from that four by four – four hours digging four hours taking action to that seven by one approach has allowed us to have a much larger footprint and visibility to what shrink is happening in our organization.

Data Management With Profitect.

So give me my data and that’s essentially what IT knows me for because I just asked for everything and if we decide we’re going to use it that’s fantastic if we don’t use it we’ve got it anyways for some sort of future correlation. You know traditional EBR is going to look at it, log data and you’re going to get one side of the picture and you’re going to be known as that cops and robbers types of LP organization that you know we probably all grew up in. But as we begin to ask for data outside of the T log,  outside of your ecomm type of data like incident reports, LP assessments, customer loyalty data, and non-receipt return data, we’re able to then span across the organization and get insights into shrink truly in real time.

Loss Prevention Analytics With Exception Reporting.

So one of the unique things that we have going for us in our organization is the empty box log. And I’ve seen my team back there and they all love the empty box log. So thanks to our non paying customers they generally leave a box behind. We don’t have RFID in our organization so I don’t know when that box is gone or empty but that empty box log feeds on a daily basis. So we’ve got this trend that we’re able to identify data anomalies, exceptions with vendors styles, items down to the color level if we need to and that data coupled with all these other data elements that we’ve been able to feed in have really allowed us to span outside of our LP function into other areas of the business.

Prescribing and Managing Actions.

So, prescriptive actions as  we talked about it as we’re looking at incident reports or assessments we require our stores to do a lot of different things on a very limited payroll budget and I’m sure everybody else is faced with that as well,  but what we’re able to do with the solution is when an exception happens, we can actually down to that user level without having to touch a regional loss prevention manager give a prescriptive action that we set as the leaders in our area to the business to take action on. And that spans outside of a store, if we’re sending it to an allocator, the allocator doesn’t need to know anything other than what we tell them to do those actions are built into the tool, and the taps management functionality has allowed us to be very prescriptive in how we are approaching shrink and loss in real time.

Flexibility & Ease of Implementation For I.T.

I also want to talk about the flexibility that Profitect has enabled us to have. So can I get a show of hands that  IT asks you to give them more data or asking you to put things on their plate? No? No. Yeah so our IT project list is a mile and a half long we can’t get anything done if it’s going to take IT resources. That’s where Profitect has really helped us because we can take the data that was traditionally going to our EBR tool, send it right over to Profitect and their data integration team is doing all the heavy lifting. You know our CIO was was very apprehensive as we first engaged in conversation with them and literally within a day or two we had our data up and running I know in the past that took weeks if not months I’ve heard years from other solutions, and getting that data up to speed that quickly, with limited IT involvement, gets IT giddy to help you in the future which is really where we’re trying to go.

So again some of the key focus areas and I’ll highlight some of these fantastic examples that we’ve had stem from the inventory data that we’ve historically had some damages and  RTV’s which have a direct correlation to number one, the bottom line, the profitability but also customer sentiment from a returns perspective. Nobody wants to buy footwear product or buy a purse or a handbag that is defective. Why wait for that defect to get to the return if we can identify it based off of a few examples versus many. Again the empty box analysis, the shrink analysis, so when we when we pull shrink twice a year we’re able to take our raw shrink data feed it into the Profitect solution and then instead of having pivot tables and reports we’re able to quickly identify anomalies in that data sometimes actually being able to reconcile the data before its financial booked which is really been a great benefit for us.

Profitect’s Inter-Organizational Impact.

The other thing that’s been able to do is help change the perception of how we’re viewed as an LP within our with our own enterprise. So we’ve now come to the point where other business partners are reaching out to us for insights into their problems and we’re becoming actively engaged in fixing operational issues that may them outside of a traditional loss prevention focus area,  just because we have the data and it’s easy to get to when we talk about POS and sales data. You know you’re looking at your traditional discounts, coupons, loyalty issues, and then other operational controls by feeding in again those LP assessments, your incident report data you’re able to identify exceptions and shrink drivers that are going to impact the bottom line. One of the most exciting things that we’ve been able to do with Profitect over the last year so is build a predictive shrink model.

Shrink Prediction,  Data Visibility, And Ease of Usage With Prescriptive Analytics.

So we’re able to now predict our shrink at a point in time with about eighty-five percent accuracy directional to where our number is going. We don’t have any other solution in our business that can give us that much surety of where our shrink is going. It’s also helped us plan budget and accrue at a better rate so at the end of the day our shareholders are actually getting value from the solution that we’ve invested in.

And I think you know data is knowledge and you know just having a solution that allows you to look at data in a million different ways puts you in the driver’s seat in your organization. You know I remember when I started my career at Nordstrom you know we would have to ask corporate for a lot of things being in the store environment. Now our end users have all of the answers that corporate has which is fantastic. And then as you think about change so again traditional exception reporting it’s that pull type of solution that you guys are used to and that we’re used to and there’s definitely a user management and a change management functionality that you have to think about as you move to a prescriptive and predictive analytics solution. Someone having grown up in loss prevention I get it I was used to digging for that data and it’s fun quite frankly but why spend the time digging when the solution empowers your users to take action on something without having to dig all of that. Allowing the system to detect the pattern so you with the solution essentially identify what condition you’re looking for and then put in whatever action you’re looking for follow-up and again that task management feature that’s built in there is going to immediately give your end user the ability to take action, resolve the problem and move on very quickly which is quite frankly hard to do if you’ve got a legacy or traditional EBR tool. I also know there’s a lot of people that are addicted to sequel or SQL it’s never been my forte you know there’s definitely users that love sequel you don’t need sequel anymore. Your sequel users as smart as they are can literally jump in this solution on day one and start finding exceptions and anomalies in your data which is to me pretty exciting.

Machine Learning & User Feedback As A Feature In Profitect’s Exception Reports.

The other thing that this solution does that in the traditional EBR tool doesn’t do is the machine learning and the end users feedback is what makes the power or makes the system powerful. So as exceptions or opportunities happen it’s up to the end users in the tool to respond within the system and say hey we need to tweak this condition. Our end users are actively,  on a daily basis, with every pattern, making edits and tweaks to those conditions in those patterns. Which then in the future generate better opportunities, eliminating false positives and that’s the key eliminating false positives again not enough time not enough resources and not enough money to really focus on all the problems that you have and I know in our 508 stores and I’m sure you guys have in your environment. And when I think about those prescribed actions those prescribed actions are dynamic they can change if a policy changes your prescribed action changes. If you want to include your policy in that prescribed action so that end-user has no question what you’re asking them to do,  PDF it and put it on it that’s exactly what you need to do within the solution that’s how we’ve empowered our end-users to take action immediately.

Loss Prevention Examples With Prescriptive Exception Reports.

So I want to talk about some of the examples and results that we’ve had with the solution I think that’s probably what you guys are here to find out. So we talked about the empty box log and Timberlands our paying customers love Timberlands and our non paying customers I think like them a little bit more and what we quickly realized last year is that this one particular color, not style,  but color of Timberlands the shoplifters were particularly fond of it and with the pattern that we had generated with our analysts we were able to identify that 82% of our shrink was only coming out of 27% of our ownership. Which gave us the ability to take action quickly and precisely to have an immediate impact on shrink in real time. The beauty of this pattern is that it was a very limited impact to the end-user. Meaning our stores and our customers did not deter sales and had about $100,000 impact to the bottom line within a 10-week period which is fantastic. We were then able to take that, generate future patterns and this actively runs in our organization. So when a style reaches a certain threshold that action is immediately taken and we can immediately move to a course of action like pulling the left which is exactly what we did with this. Again our environment is open sell,  we don’t want to take the customer or the product off of the floor, away from the customer, but at the end of the day there was nothing we can do – a tag, a person, or any other technology, other than pull that left to have an impact like we had here.

We did the same thing with other, unfortunately, UGG is again something that Profitect was able to identify based on specific, styles,  items, and colors we were losing at a rate way too high. Damages and RTVs so again we like to believe we sell great product and I believe we do but sometimes things happen we’ve got product quality issues that happen just like all of us. (Referring to slide) This number down on the bottom is really small but it’s 1.3 million dollars. That’s 1.3 million dollars of opportunity that I have one analyst, one end-user able to impact in our organization. I don’t have a team,  I’ve got one person that mines this opportunity in this pattern that looks for styles that have a damage higher than the subclass looking across the entire organization, across styles, items, color levels, we’re able to use that benchmark data to find exceptions and take action almost immediately. So as you guys are looking at this and I’ve blurred out the vendors because of course, I don’t want you to see that we’re looking at. Both in season and prior season data, things that we can take action on immediately because we’ve got inventory actively in our source and things that we’re going to go back to that vendor and say “Hey you sold us something bad we need future buy credit.” that 1.3 million dollars continues to generate back to the bottom line and that’s what created us to be really looked at as a prophet hub for our organization again. I’m sure you guys would all love to have teams of analysts looking at this stuff but one user can generate that type of value and has generated that type of value for our organization. And it will continue to generate value as long as we sell product.

And this is a really exciting story so as I mentioned week 1 of May we rolled out the point-of-sale in the ecomm data and I remember vividly being in the training with our investigators and our end users when literally two hours into the training we found a case for eighteen thousand five hundred dollars worth of product that was shipped,  and it’s a little bit embarrassing that we didn’t have that visibility to the data before, but I think we’ve all been in that boat where reports, reports, reports, and that’s what we were doing before. We were looking at reports trying to find this type of exception literally within two hours our end user built a pattern that looked for an exception. They were looking at you know lost and stolen orders that were shipped 50% higher than the benchmark that means that anybody else in the environment in that call center they’re 50% worse and you can tweak that number a little bit. You can look at 40% 30% but 50% for us is really where our starting point was. Again $18,500 worth of orders shipped to this individual to her friends, marked down for a variety of different reasons we were able to quickly resolve that literally days after this training and then this pattern is generating continuous value.

In fact, we just closed out about two more cases last week, very similar instances but instead of waiting for weeks we’re waiting the next day when the data feeds through our system. Which is awesome, so I think you know this is this is really the true power of the prescriptive analytics where that exception across a large population of data with very limited end-user involvement has empowered our users to take action in a very limited amount of time. And I think the other thing to mention is as we rolled out the sales module we’ve had other exceptions. Nothing like $18,500 but very quick wins as we rolled out within weeks of having our data we’re learning things and I know I’m going to take – you know take away some of the questions but we’re learning things insights into our data that we didn’t have traditionally. I know Guy and I will talk about that in the Q&A; section but you know the benchmarks for us is really what helps us and our end users identify things that you wouldn’t see if you were just looking at a single store, or a single, district or a single region.  And again limited resources, one analyst, five investigators, five RLPMs, 508 stores, plus you know a multi-million dollar ecomm online business all of this data into one solution, coupled with the flexibility of the partnership that we have with Profitect has allowed us to really change the landscape in our environment. You know I think the most important thing that I would advise and coach you know you guys being my colleagues on it is really when you’re thinking about how you want to be viewed we’re not going to catch our way to low shrink. I think that’s how we all grew up and probably thought we could you know catch a shoplifter here and there but this $18,500 example or the 1.3 million dollar example is going to be way more than any amount of shoplifters anybody can catch. So our investment in this solution is long-term again since 2014 we believe in the tool, the tools helped us look differently through our organization and our shareholders which is what’s important. Guy?

Q&A SECTION

GUY:

Thanks, Jordan time for some question so, first of all, great presentation.

Jeremy:

Thank you.

Guy:

Thank you, Jordan. You know when I’m looking at what you’ve done what you’ve done so far and all great work, but your you know your unique loss prevention area you’re looking at you spoke about (and) cases that you found you spoke about, dollars that you generated, you were speaking about damages, can you tell us a little bit about whether you report into, and how do you look at total loss and loss you know what  are the measures that you look at?

Jordan:

Sure yeah, so I have the unique privilege of reporting to the CFO now. So historically we had reported through operations which was great but reporting to the CFO you know  puts a whole new outlook on how we have to approach loss and it’s not just about shrink, shrink is obviously probably the largest portion of our job and my job, but it’s also challenged us to look this holistically. At profitability and how we impact the bottom line for our business.We can’t just focus on catching bad guys we have to focus on making improvements to the business both operationally from an investigative perspective, as well as you know back to the bottom line.

GUY:

Which is – which is great and unique at the same time.  You know if we’re speaking about some you know push to pull there’s a lot of questions that we signed up with because we have about five minutes and I do want to open the floor for some questions. You know one of the things that really matters is about —  you mentioned true and false positives you, mentioned seven-by-one and four-by-four 7×1 and then you touched upon other organizations that are using your solution. What are the other organizations that are using it and how does it gain more value to loss prevention?  Most of our people in the audience are loss prevention, why should they care about engaging other departments with the same — you know the same application?

JORDAN:

Yeah , again I think you know historically data had been in silos for our organization,  and as we implemented the Profitect solution and started sharing some of the the highlights through a Profitect steering committee that we started with executive sponsorship,  that visibility that we created to our data, created a drive and a need for other people to ask those questions about their data, and you know again we have one analyst. We don’t have a team of analysts like other units of the business, so we were able to really kind of inspire other parts of the business to want to interact with us and ask, for — for help and the solution is flexible enough that you don’t just have loss prevention engaged, but you’ve got merge planning analysis, advanced analytics, store operations, inventory control, and everybody’s actively using the system in some sort of different capacity.

GUY:

Good, very good yeah sure that there will be any other specific question about it please come over but if you have a question we only have about five minutes. Please raise your hand and someone will come and get your question. So you know you mentioned your CIO support, you mentioned your IT,  you know, how did you get them so you know that every retailer that I am speaking with they have tons of IT projects, as a matter of fact we are actually going to NRF Protect in July and we’re going to meet a lot of CIO’s. Every CIO in retail today is doing more with less, outsourcing their IT department, changing point-of-sale, the e-commerce, unified commerce and a lot of other things. How did you get your IT attention to you know to help you do your overall —  you know changing your EBR status?

Jordan:

Yeah so I think it starts with number one being a good steward of the business and understanding the difference between EBR, prescriptive analytics,  and what other solutions are out there. As I approached my CIO and our IT leadership team it was about the things that Profitect could do that the other solutions that we either had or were going to have can’t do. We’re implementing MicroStrategy right now as the BI tool, Profitect is not the same thing as a BI tool, and I had to do some education to my CIO, which was an extremely difficult conversation and Profitect helped you know. You guys helped absolutely, the whole way with that education about the difference between EBR, BI, and a prescriptive analytics solution.

GUY:

Yeah a lot of times. They say you know,  it’s not another BI tool, right? It’s not replacing a BI platform but in a simple way any report that should generate an action you should just tell the person what action they need to do.

JORDAN:

Yeah we just — we just ran through a project a data governance project and we had over 1700 reports and I think the read rate was five and a half percent. So that just goes to show you I mean you can have reports to until you’re blue in the face, but if you have nobody opening up those reports and even if they do open it they don’t know what to do, that’s how we were able to educate the business, specifically our IT leadership on the true difference between a BI tool that can visualize something, and a tool that’s going to inspire action.

GUY:

Good so another, you know, another question about what surprised you the most from a prescriptive analytics solution perspective —  after you went live with the knew EBR solution?

JORDAN:

I think we’ve had multiple aha! Moments. You know what was unique about how we transitioned from our former EBR vendor to Profitect is we actually send them the same exact data. We didn’t reinvent the wheel we took the same exact feed and that’s how we moved so quickly. Duplicated the feed and sent it right to you guys and what we quickly learned is that our legacy solution was missing correlations and missing the identification of shrink that was happening, or loss that was happening, and I was blown away by how quickly; number one, we got it up, but number two how quickly were able to get insights into our data that we actually should have had before.

GUY:

So you know you spoke about the change management, another probably curiosity in the crowd is around the push to pull model. You probably had some people that are addicted to sequel and well you know analysis paralysis so some of —  one of our customers used to call it hotel analysis. You know you go to your stores, you’re sitting in your hotel and you’re doing your analysis before you go to the store. What was the discussion with your team and how did the team actually accept the push versus pull model?

JORDAN:

It was — it was a dramatic change from what we were historically used to. You know the generality is or the reality was that we were that type of hotel approach and we worked with our users and through our users to help them understand why we were making this transition from something that just like retail, was working one way and quite frankly today it’s not working. And that’s how important it was for us to make that change and get our users bought in, and it’s not easy there’s definitely people that are still out there today that we’re working with,  making sure that they’re in the tool, they’re using it all the ways its intended to be used, but the task management feature — the accountability down to the end user helps us manage something before it’s a problem.

GUY:

Guys any questions?  We have probably another minute, we’re running out of time. Anyone –  any questions, no? Yeah I would like to thank Jordan for joining.

JORDAN:

Thank you, Guy, I appreciate it.