On the eve of the NCAA Final Four basketball tournament, the news show “CBS Saturday Morning” did a feature story on analytics in sports. There were some great lessons in the piece that not everyone has applied to casino operations.
Naturally, the feature began with references to “Money Ball.” That was the hit movie starring Brad Pitt and Jonah Hill about Oakland A’s manager Billy Bean and his “data analyst” Peter Brand. It’s a wonderful film that was nominated for six Academy Awards. While there are differences, the 2003 book by Michael Lewis (the inspiration for the film) is even better…and very instructive. A NY Times reviewer said, “The bottom line: in the American League West last year, the teams finished in inverse order to their payrolls. Oakland wound up ranking highest with the least money, demonstrating a principle that can be appreciated far beyond the realm of baseball.” If you haven’t read the book or seen the movie, you should.
The basic plot of both is that you could build a better baseball team if you used analytics to select players instead of the traditional “gut feel” from scouting reports and/or fan favorites. That one concept alone changed baseball. The Boston Red Sox used it to win their first pennant after a drought that lasted 86 years (or maybe it was just the curse of the Bambino that lasted from 1918 to 2004).
Those lessons were not lost on progressive slot directors who, about the same time, started relying more on analytics to decide which players (machines) belonged on their floor, rather than relying on sales literature, G2E cocktail parties and salesperson charm.
Indeed, the 2020 pandemic year elevated respect for slot data analytics even more, as many casinos generated more money with fewer machines during restrictions by applying their analytics even more aggressively.
However, the CBS News story talked more about operational analytics than just player (or slot) selection. They cited the fact that prior to analytics, football coaches went for fourth down conversion less than 1 in 20 times. With analytics, that number has dropped to 1 in 4 attempts.
Likewise, analytics changed the game of basketball when the data showed the value of the three-point shot. Both the Houston Rockets and the Golden State Warriors used those stat-driven lessons to prioritize players and plays designed to hit the long-range shot. Both teams who began using these strategies won regional and world championships before others changed too.
While most casino operators now use some form of analytics to improve individual slot selection, only a few are using their analytics to improve their operational strategies.
The rebound from COVID staff reductions and operating restrictions is an excellent time to evaluate (or re-evaluate) your team’s operational strengths and weaknesses. Do you have enough team members … or too many? Are they working the right shifts? And in the right places at the right times? Are they all efficient? The answers are in your data.
One of the first companies to mine this resource was CIS Technologies out of Reno/Las Vegas. Their early “slot monitoring” application was originally designed as a dispatching system to manage timely responses to change lights, jackpots and tilts. But the developers quickly realized that there was a trove of information available that could benefit slot operations in a variety of ways.
CIS gathered basic data from your existing slot system and analyzed it against similar numbers from your HR/Payroll systems. In other words, they could contrast slot floor activity against team members. Add in days and time-of-day, and the results for most users were a revelation.
The graph illustrated here plots a 24-hour day from a monthly period. The green bars represent the number of slot events (Change Lights, Jackpots and Tilts). The red bars indicate service times in minutes (i.e. how long it took to respond to a task). The dark blue line shows the number of slot team members assigned to the floor. Importantly, the light blue line is the number of team members actively working on a service call.
This graph is from a casino which had used this application for several months and had made major adjustments to their operations. In an ideal world, three of the metrics (with their scales adjusted) should show similar slopes and curves aligned with one another. The exception are the red bars, which should remain flat, and hopefully are as low as possible (constrained by labor budgets).
The peaks in the dark blue lines represent normal “shift” changes and time set aside for “pre-shift briefings.” Otherwise, the small gaps between light blue and dark blue are normal, since it is unrealistic that team members would on service requests 100% of the time. This chart is nearly ideal, except for some work that could be done to improve service times at midnight, 2 a.m. and 7 a.m.
According to Jim Eller of CIS, “Using a contrast or ‘before and after’ approach via one of our reporting tools, we help a property understand where the opportunities to improve are within the staffing model. “
“Within the first 90 days of implementation and using the training on how to interpret what they are seeing, about 75% of the properties have reallocated the staffing roster toward the ‘customer service mountain’ they climb daily. This in turn creates a better experience on the slot floor.”
In other words, most graphs they see among new users show the different metrics peaking at dramatically different times. Getting them in synch can make a huge difference in productivity.
The same methodology also applies when analyzing the effectiveness of the technicians on the slot repair team. Eller says, “When a property can use analytics to create pre-emptive game repair workflow and combine it with the shortest possible wait times for the guest, they get high marks and positive feedback from both guests and team members. Believe it or not, effective tools and training can also improve a slot operation team’s moral toward the guest experience.”
Today, many standard slot systems have some of these floor monitoring features built in. If yours does not, there are excellent tools for those with the tech skills to DIY their data with apps such as Tableau, Agilisys and others. (Author’s Note: I think the CIS solution is still the best for this purpose).
Graphs like the one shown can be weekly/monthly/quarterly; or drilled down to individual days-of-the-week. They can also be concentrated on specific zones or even varied job descriptions.
“Looking at response and completion times for slot operations, beverage service, casino host & player development, slot technical repairs and carded tier-level programs,” Eller explains, “all play a part in the total experience of the guest visit. Examining event flow (such as jackpots, machine tilts and service lights) rather than just coin in and head count which then determines staffing requirements is a new paradigm for most properties.”
A controversial feature of such monitoring programs is their ability to measure the response times of individual team members to certain events. While it may seem logical to offer rewards for those with fast responses and discipline for those who are slower, that can be a mistake if not done judiciously.
For example, discouraging slot team members from “talking too much with players” to improve service times can be a big mistake. Socialization is one of the strongest motivations for casino repeat visits. Attentive listening to the details of a player’s gall bladder operation or admiring the latest photos of their grandchildren is an important and productive part of the job.
Accordingly, any compilation of service times by individuals should only be done over longer periods, ideally no less than quarterly. With this caveat, the comparison of service times can be very meaningful. It is especially valuable if you work with your “slowest” team members to determine where and why they are inefficient.
Along this line, some overlooked KPIs are the individual components within any given task. When was the last time you measured these? In slots it could be “jackpot payoffs.” Typical J/P components include: 1. Time from a JP event to the arrival of a team member (this is a standard metric that is available by most slot systems). 2. How long before the JP was entered into the system? (do you have enough system terminals in the right locations?) 3. How long does it take to complete the required paperwork such as W2Gs and CTRCs (write your Congressperson immediately and urge them to raise the taxable level from $1,200 to a more sensible $5,000!). 4. How long does it take to get the required funds from the cage cashier (see below)? And finally, 5. How long does it take to make the final payoff and return the machine to service (are your banks conveniently located)?
Time #1 is a staffing issue combined with proper zone allocation. The period from the competition of measurement #1 to start of #2 is generally the guest interaction component of socialization mentioned above. However, with that one exception, most of these times are operational issues that you can, and should, address.
Number 4 is a common issue in many casinos. To achieve cashier efficiency, often the same cashier that supplies jackpot payoffs handles end-of-shift “bank-outs” for F&B and table games. That is efficient for the Cage but almost always means longer wait times for slot players.
If you really believe that training, or a memo, will convince a cashier to handle a jackpot from a slot person (with no tip potential) ahead of a blackjack dealer or cocktail waitress (they tip generously), you are living in a fantasy zone ZIP code. If slot payouts and other banking functions are separated, your response times and profitability will improve.
While socialization with guests is critical, the same does not apply to excessive chats between team members in back-of-the-house areas. Examining times between #4 and #5 can indicate problems, especially if contrasting individual team member’s times against averages.
Another benefit of analyzing slot floor activity by “time-of-day” is it can provide important operational data for other departments. It is not uncommon for the hours of operations of secondary services like restaurants and gift shops to be out-of-line with floor volumes.
If you look at the graph above, you can see that the busiest time of the day at this casino (a blend of all days of the week) for slots is from 11 a.m. to 2 a.m. Yet this casino, before utilizing CIS, operated both their gift shops from 8 a.m. until 6 p.m. Clearly, there was more business between 6 p.m. and 4 a.m. than between 8 a.m. and 10 a.m. Floor activity peaked at 10 p.m. and stayed strong for several hours when the gift shops were both closed. That was changed and (surprise) gift shop sales increased, as did guest satisfaction. These same metrics can be especially useful in setting hours for cocktail service, cabaret performances, host staffing, etc.
Another source of slot player dissatisfaction is a faulty card reader. If a bill, ticket or loyalty card is “rejected,” complaints will follow. Lost tracking reduces effectiveness of marketing, bad bill validators can ruin a machine’s performance, and it’s extremely frustrating to insert a card, tickets or bills multiple times in an attempt to get a “green light.” All slot systems record these “reject” events, but few offer good reports to measure them. It is easy for your IT department to create a list ranked by severity. Thus, techs can prioritize fixing bad readers. It will make an immediate improvement on your guest satisfaction surveys.
Again, if you are not utilizing operational analytics, you should.
