Thursday, October 6, 2016

Hi Chefs - Use Culinary Data Science to Optimize Your Recipes and Menus

(Click on the image to enlarge)

Reader's Review
I just finished reading the book and I just wanted to let you know that I really enjoyed it.  I love the approach you took with it.  You were able to give all that amazing info and analysis but by telling the story you did it in a way that is easy reading and very compelling.  There’s a lot of valuable info packed in here that anyone who has an interest in restaurants needs to be aware of.  I was hoping for a different ending but I’ll get over it.  Well done Sid and I hope ChefQuant takes off!

Link to the Book on Amazon

Our LinkedIn Chefs and Restaurant Owners/Managers may request Complimentary PDF copy of the book!
Chefquant@gmail.com

Monday, October 3, 2016

Announcing ChefQuant – A Practical Guide to Data Analysis for Restaurant Owners and Chefs

Press Release

https://www.prlog.org/12591504-announcing-chefquant-practical-guide-to-data-analysis-for-restaurant-owners-and-chefs.html

ORLANDO, Fla. - Oct. 3, 2016 - PRLog -- Sid Som, the author of popular info sites like HomeQuant,  Homeyada, TownAnalyst and SkylineValue, is pleased to announce the availability of his newest e-book/paperback, ChefQuant, on Amazon.com. It's a practical guide to data analysis and modeling for restaurant owners, managers and chefs.

This book will teach the readers the issues and techniques that are rarely found in textbooks and courses offered by professional institutions. It covers such hands-on chapters as location planning, demographic analysis, menu planning, pairing wines, sales analysis and rewards promotion, profitability analysis, seasonality to year-round business, alternative capital planning and real estate strategies, re-steadying business plans, etc. to help make more informed decisions. While this book is intended for restaurant owners and executive chefs, it could be very useful for anyone planning to start a non-dining retail business as well. People starting out as general financial and statistical analysts could make use of it as a primer, while business and economics professors can recommend it for the case studies.

While some restaurant review sites have gained tremendous popularity amongst consumers and media folks, they do not necessarily help the restaurant owners in making any meaningful operational or strategic decisions as those surveys do not constitute scientific polling, thus the underlying atomic data are generally unsuitable for any consequential statistical analysis or econometric modeling. Additionally, such surveys are usually few and far between for the smaller outfits and start-ups.

The primary objective of this book is to educate the owners and managers of independent restaurants as to how to apply a set of easy to use yet highly scientific methods to analyze and model the right type and amount of data. Given the highly diverse nature of restaurants and the general ease of data collection on their part, this book therefore focuses more on analyzing and modeling data and forecasting outcomes than on standard data collection methods, with the exception of specialized events like demographics, location and seasonality.

Data collection is generally a simpler process for restaurants than many other types of businesses. Each time the cash register rings, an atomic sales data is created. Of course, sales data alone does not provide a sound basis for technical analysis. They have to be properly complemented with the satisfaction data. A few straight forward survey questions immediately following the sales provide the feedback loop leading to enormously powerful and minable technical data. Then, by meaningfully connecting and mining the combined sales and satisfaction data, independent establishments may arm themselves with the knowledge that is utterly germane to informed and intelligent decisions.

Instead of writing the book in the form of a traditional text book, the author used a storyline to communicate with the readers, thus making it a fun experience as a whole, rather than forcing them into a black box and asking them to figure things out.To make the experience even more palatable, all mathematical equations and formulas have been steered clear of. While this is not a science fiction, this book additionally proves that science and fiction could be seamlessly married for maximum synergy.


Saturday, October 1, 2016

ChefQuant - A Practical Guide to Data Analysis for Restaurant Owners, Managers & Chefs

(Click on the image to enlarger)
Read it for the science. Read it for the fiction. Worth it, either way.
Link to the Book on Amazon

Available in Both: Kindle/Device and Paperback    (Search: Chefquant Book)

Reader's Review
I just finished reading the book and I just wanted to let you know that I really enjoyed it.  I love the approach you took with it.  You were able to give all that amazing info and analysis but by telling the story you did it in a way that is easy reading and very compelling.  There’s a lot of valuable info packed in here that anyone who has an interest in restaurants needs to be aware of.  I was hoping for a different ending but I’ll get over it.  Well done Sid and I hope ChefQuant takes off!

Free Reading:
Chapter 1 
Chapter 2

Monday, August 1, 2016

Press Release: Launching ChefQuant – A Restaurant-level Data Analytics and Quantitative Solutions Marketplace

PRESS RELEASE
http://prlog.org/12576600

ORLANDO, Fla. - July 31, 2016 - PRLog -- Sid Som, owner and publisher of prominent data analytics sites like Homequant, TownAnalyst and LocValu, is pleased to announce the launch of ChefQuant, an online data analytics and quantitative solutions marketplace for restaurant owners, managers and chefs. Unlike trade organizations that focus on macro industry trends and generic industry solutions, ChefQuant will offer custom solutions to individual restaurants, effectively mining and modeling their internal and external customer survey data. A trend-setting book 'ChefQuant - A Practical Guide to Data Analysis and Modeling for Restaurant Owners, Managers and Executive Chefs' spearheads the launch.

While the vast majority of non-chain restaurants have access to their customer data, they do not necessarily possess sound know-how to explore and mine their own data reservoir. Sid contends each restaurant has its own unique micro market, often significantly bucking its (macro) industry trends and nuances. A restaurant's data reservoir is therefore its own internal goldmine with hidden answers and solutions waiting to be harnessed for their success and profitability. Industry-level data solutions are too broad-based to be plug-and-play for all; consequently, one-size-fits-

all industry solutions often require costly re-engineering to be, at best, somewhat productive for most members. Instead, if they are provided some hand-holding in understanding and modeling their own data, they will perhaps promptly learn to address their own shortcomings and remain on a profitable path.

The reason Sid started his Blog ahead of opening the actual online marketplace is to demonstrate to the target audience how his proposed solutions could be of tremendous help for all restaurant owners – successful or not-so-successful. Even those who are contemplating opening new restaurants or adding new locations can benefit immensely from his solutions. Since ChefQuant solutions are based on a restaurant's own data, rather than borrowed generic data from the competition or the industry as a whole, they are totally custom tailored addressing their own unique issues and challenges, hence incredibly result-oriented.

Depending on the individual needs, ChefQuant will provide solutions ranging from demographic analysis for new locations to customizing menus for generational tastes to identifying deficiencies in existing menus to aligning entrées with adult beverages to forecasting sales to developing internal performance and service metrics, to name just a few possibilities. Needless to say, ChefQuant uses sound statistical, econometric and operations research techniques in mining and modeling all data.

In a recent conversation Sid mentioned, "The non-chain restaurant market is one of the most competitive markets out there. One of the reasons many succumb to the competition is that they do not attach much importance to data analysis and modeling, which in most cases hold the key to their success or failure. Granted, restaurant owners, professional managers and chefs are not expected to be trained quantitative technicians, which obviously created a dire need for external professionals to step in and work with the industry to develop custom solutions. I urge all restaurant owners, managers and chefs to read my book  and decide for themselves if we make sense. Bring your data to us and we will show you the path to continued success and profitability."

Thursday, July 28, 2016

Modeling Internal Customer Survey Data to Identify Inconsistencies & Deficiencies in Menus


Chapter 2 (from the Book)

Modeling Internal Customer Survey Data to Identify Inconsistencies & Deficiencies in Menus 


Despite having easy access to the customer response surveys, most non-chain dining establishments do not collect any such data. Even when they do, they rarely spend any time and money mining the data; instead, they depend heavily on macro industry data and research in combination with some unscientific samples derived from anecdotal opinions and responses.

Obviously, random business decisions, including menu planning, based on such general solutions tend to produce hit or miss results, often adversely impacting customer acquisition and retention, lessening the prospect of long-term growth and expansion. Hopefully, the new generation of data savvy entrepreneurs will look more closely into their own internal data to decide on competitive strategy, absent which they will risk being part of the strikingly high small business failure statistics.   

Two days after the first business meeting with John and Priya, ChefQuant was pleasantly surprised to see an envelope on his desk with a handwritten note, “Please model it out.” As he looked at the content of the media on his computer he realized it was their internal customer survey data that Priya had seemingly started to research. The timing couldn’t be any better as ChefQuant was looking for a new project for his new batch of interns. A veteran intern was given the lead to indoctrinate the new batch into their new world of culinary data science.

ChefQuant was proud to see that under the able leadership of the veteran intern the new team developed the initial model in two days. He was taken aback by certain findings so he immediately called John and set up a meeting to share the modeling results. The following morning as John and Priya showed up ChefQuant decided to explain their research methodology before getting into the actual results.

ChefQuant explained, “First off, thanks to Priya for providing us the data in soft format so we didn’t have to do the data entry. Although she has given us six months’ worth of data, our initial analysis of the data pointed us to its consistency, thus encouraging us to work out of a random sample of 100 cases instead of the entire population itself. FYI, a quality representative sample is as good as the population.”

Priya was curious, “What’s the difference between a quality sample and a perfect sample?”

ChefQuant pointed out, “There is no such thing as a perfect sample. A perfect sample is the population only. A quality sample, on the other hand, is one that is statistically significant, hence heavily relied on in data modeling. That is why most Stat or Econ professors generally avoid using the term ‘representative sample’ because a sample, by definition, is representative of the population. Of course, our clients are mostly business owners so we take a more colloquial approach in our presentations. Now that you have an understanding of the methodology, let me start with this simple frequency table that summarizes the survey data.”


(Click on the graphic to enlarge)

ChefQuant continues, “Given your restaurant’s popularity and positioning, I was not too happy to see scores below 4. While this table doesn’t show the median, the median score checked in at 4.50, with 19% customers rating the overall experience as perfect. Obviously, there is significant room for improvement, which the following correlation matrix demonstrates.


(Click on the graphic to enlarge)

The overall score (“Score”) and your offerings must be collinear. In fact, the scores in all of the individual categories must always be above .85, ideally above .90. Honestly, it is shocking that Beverage and Dessert categories are hovering around .60. Going by our internal metrics, I would reason that the Wine List you are offering is not properly complementing your entrées while the formulary dessert combinations are not adequately appreciated or promoted. In your type of environment, the correlations between entrées and other menu items must be above .60 and even higher for presentation. The low correlation between Dessert and Presentation is clearly another red flag.


(Click on the Graphic to enlarge)

Now let’s move on to the regression model. The reasonably high predictive relation between Score and the other variables, as evidenced by the correlations matrix, paves the way for a reliable regression model. In this model, Score is called the dependent variable while the other variables are called independent variables. These independent variables, collectively, help explain the variations in the dependent variable. Unlike the correlations matrix which depicts the existing relationships, this is a predictive model that evaluates all of the independent variables as a group.

Even the model is revealing that Beverage is the weakest candidate in the group, with the weakest coefficient and T-Stats and high P-value. Conversely, Appetizer is the MVP with the strongest coefficient and T-Stats along with low P-value. The reason is quite simple: Your customers come to a place like yours for a memorable experience which obviously starts with the appetizers. Amazingly, the model is predicting that Dessert could be the next best candidate in the lineup by projecting its high T and low P, but it’s not there yet as it carries a weaker coefficient than those of Presentation and Service. The fact that Entrée hasn’t fared well is self-fulfilling in that it needs to be better coordinated with Beverage and Dessert.


(Click on the graphic to enlarge)

Finally, take a look at the above scorecard presenting the first few responses from the sample. Remember, once you have the modeled scores, you must use them in optimizing any process or menu as they are statistically more significant, with reduced judgments that are inherent in all opinion surveys. For example, the first respondent gave you a perfect 5, though two categories contain straight 4s, whereas the model returned a more meaningful 4.8. Again, consider the last but one where you received an overall score of 3.50, while the model revised it up to 3.75. Well, that’s all I have on this. Any questions?”

John asked, “So, what’s the fix, ChefQuant? Where do I start?”

ChefQuant’s answer was quite simple, “Share the analysis with your beverage manager and try to understand where the disconnect is. If that doesn’t help, I will hook you up with a wine consultant friend of mine who specializes in coupling wines with entrées. He is very good at it. Then, after your Chef leaves, let Priya spend some time with your Pastry Chef and figure out the underlying issues. If the current dessert menu lacks proper flow-through, overhaul it. Also, do some role playing with your service staff to making sure they are using the right pitch in promoting the current dessert menu.”

John remarked, “I have to get something off of my chest. Now I know why my banker repeatedly asked me to consult with you, ChefQuant. This is the kind of class I had in mind when I had signed up for the Applied Finance class in my undergraduate. Instead, the professor spent an entire session manipulating an alpha which I could hardly see from my seat. I never knew applied science could be so much of fun, yet so incredibly meaningful.”

Sunday, July 17, 2016

Modeling the Market Survey Data to Plan, Develop and Optimize Restaurant Menus

Chapter 1 (from the Book)


Why Demographic Analysis is so Critically Important in Planning Locations and Menus 



While the vast majority of large restaurant chains employ qualified market researchers and quantitative analysts, working with in-house Chefs and testers to plan, develop, and optimize menus, most non-chain and smaller restaurant chains are generally financially constrained from taking that robust research route, and instead rely, at best, on focus groups and internal customer surveys to tweak their menus.

Most focus groups are far less consequential than statistically significant market surveys, while internal customer surveys require sophisticated data tabulation and mining, often packing them to collect dusts. For example, a mere 30 responses a day over a 60-day period will require the help of a data analyst and statistical software to tabulate, analyze, and model the data. If the menus are seasonal, the data collection must be repeated across seasons, necessitating even more sophistication.

Many successful non-chain restaurants located in busy tourist areas or demographically cross-sectional neighborhoods suddenly face disappointment when they try to duplicate that successful “central” format in other parts of the city with vastly different demographics.

John Doe is a very successful restaurateur owning and operating a sit-down lunch and dinner restaurant in a touristy city center. He and his banker agree that the success of this restaurant should be duplicated in other parts of the city, so John decides to open two more – one in the vibrant eastside across from a large State College campus and another in the thriving westside in close proximity to a wealthy retirement community. He has seen success and hence believes in his proven format, including holding his current menu constant for the new locations.

Suddenly, a bolt from the blue.

His Chief Chef is leaving. Due to family reasons, he has decided to return to the west coast. John Doe is nonetheless adamant that he will not abandon his expansion plans. He contacts his banker friend, who suggests ChefQuant Consulting Services. While he is not too thrilled about the suggestion, considering he knows his business inside and out, he reluctantly stops by their office and meets ChefQuant to start the dialogue.  Knowing John is looking to fill the vacancy, ChefQuant invites Priya, a recent graduate at the top of her class from the local culinary school. Before the interview, ChefQuant explains to John that Priya belongs to a new generation who is not only culinary trained but is also data smart, although she lacks the experience of heading a restaurant yet. John is shocked (he was looking for a well-known Chef with at least ten years of culinary and management experience).

John returns to office and speaks to his banker friend who encourages him to take ChefQuant’s advice very seriously. Under a bit of pressure and a feeling of despair, John decides to hire Priya as his Head Chef. While Priya will be working out of the central location, she will be running the two new locations with resident Assistant Chefs and monitoring their operations via live video feeds.

Priya comes on board in a week and starts working with the departing Chef.

A week later, ChefQuant holds his first business meeting with John and Priya. Right at the outset, John makes it adamantly clear that he would not change a thing when it comes to his successful menu and even the interior décor for his new locations.  While ChefQuant did not have any specific survey data for John’s new locations, he puts forth the following citywide survey summary (median) broken down by age groups, politely emphasizing that the eastside location will primarily cater to the Gen-Y (college students) while the westside location will attract mostly Baby Boomers and Seniors 70+.


(Click on the image to enlarge)

The summary shows these two groups’ tastes and preferences are quite different from the overall medians that his central location is perhaps relying on. ChefQuant therefore emphasizes to John that the existing menu has to be meaningfully tweaked in line with the actual survey data and the resulting analyses, with the possibility of several menu items being adjusted, changed or even swapped.

John looks at Priya and wants her suggestion. She says, “Given the divergence in the data, I totally agree with ChefQuant. In fact, once the actual east and west survey data and analyses come in, I suspect they will show even more polarized results than what I see in this citywide summary, potentially prompting significant changes to our central menu. The Gen-Y’s taste for salt and sweet is noticeably higher than the overall medians, though they don’t care for bitter at all. Conversely, BBs and Seniors prefer lower salt and sugar, but are unafraid of bitter.  For example, the appetizer sampler for the west may require a swap like roasted Chinese Bitter Melon with tangy garlic sauce while the east may enjoy a salty and spicy Indian Samosas with sweet Middle Eastern date sauce.”

Priya continues, “By the way, I have been tabulating the recent internal customer surveys for this location in order to understand if there is any disconnect between what we offer and what our customers are truly interested in. I am seeing certain gaps I need to gradually fill in. The culinary business is not all-art anymore. It now requires a good dose of science. For example, this simple correlations matrix tells me how to couple a main ingredient with a complementary sauce.”

John then concludes the meeting by sharing an episode from his college days, “I was a good student, majoring in business. I was so confident of myself that in the final semester I proudly signed up for a class called Applied Finance. To make the long story short, after the third class I had to drop out like a wet cat. The class was nothing but complicated math and the freaking professor was from the Math Dept. I was wondering if I’m looping back into a similar situation, again?”

They will meet in a month after ChefQuant’s marketing team collects and analyzes the targeted survey data.
____________________________________________________________