We continue our series of articles about personalized cigar recommendations and what we consider to be the key factors for a personalized service. Today we’ll focus on product knowledge aspects and will also share with you some of the results of our latest customer satisfaction survey…
Science vs. Pragmatism
If you have read us before, you know that we operate as a lean-agile team. In alignment with our continuous improvement remit, we recently conducted a non-scientific survey among our members. The survey was meant to understand where we’re doing well and where we can further improve our service. The areas we selected are the ones that we consider to be key success factors for our personalized service:
- User Control
- Sales Motive
- Cigar Knowledge
- Member Knowledge
We thank our members for the terrific feedback we received on all these fronts.
Today, we’ll focus on the cigar knowledge aspect.
On the knowledge of cigars for recommender systems
It’s difficult to recommend a cigar if you don’t know a certain degree of attributes of such cigar. Just like it is frustrating for knowledgeable customers to get recommendations by people who don’t go beyond a fairly narrow range of descriptors such as “med-bodied” or “full-bodied”.
As explained in a previous article about accuracy in personalized recommendations, the more a website knows about cigars, its customers and their interactions, the higher is the quality in its cigar recommendations. You could say that it is sufficient to put the user preferences and the cigar attributes in structured data sets. Yes, it is as simple as that. And as complex as that.
Cigars are not vacuum-cleaners. You can construct your product descriptions based on manufacturer information and you will have anagraphic information for the cigar. Assuming you don’t want to use marketing copy for product description, as it normally does not include a full sensory analysis, you are left with attributes such as size, shape, brand, country of origin, blend, nicotine strength, price. You can use these to match your customer’s preferences if these are expressed in such terms, they are important preferences.
However – since 87% of consumers state that flavor is the most important criterion, and 9% state it is very important for them to choose a new cigar – the attributes above may be short of something.
Flavors attributes can be a bargain to get: the knowledge of the crowds is time and cost effective. Crowd-sourcing enables users to submit whatever comes off the top of their heads. They may not know much about how to identify flavors in a cigar and may not take the time to learn about it. Any naming criteria you stated may be ignored, or simply mis-used.
The best problem you can have when you want to crowd-source your cigar descriptors in view of recommending cigars is data sparsity. When you have very few cells that you are trying to match to user preferences, the worst thing that can happen is that you cannot find any new match of cigars to recommend to the user. Too bad, but it’s indeed the best problem you can have, as the contrary, when data is too dense, leads to even bigger dissatisfaction. Let’s see an example:
How do we deal with “cigar knowledge” at Cigar Sense?
We add what is required to meet the wants of those 87%+ consumers for whom flavor is the most important attribute.
Flavor and anagraphic information in our database account for over 150 attributes for each cigar. Each cigar profile is updated after having consolidated data entered by trained tasters. Training is the key. Tasters need not only to know how to describe a cigar, but also need to calibrate their nose to the rest of the panel noses.
Here is the Bolivar Belicosos Finos per our panel:
You can see a clear difference, which resides in the predominant attributes being represented in this second chart. This is what allows us to match the cigar attributes to our members’ preferences in the most appropriate way we have been able to find so far. You may read about pro’s and con’s of the use of panels in the taste-tests of cigars. We certainly know that cigars themselves imply variability because they are hand-made products resulting from the work of more than 200 people. In spite of this, we are still strongly convinced that a panel provides the most objective way to describe a cigar.
As a consequence, we believe that we will not change our process in the near future.
How do our members rate us on cigar knowledge?
We asked our members to state how much they agree to the statement:
“Cigar Sense has deep knowledge of the cigars it recommends”
We are very grateful to our members for recognizing what we strive for. Their overall score is 9.5 out of 10 for cigar knowledge.
Cigar knowledge can have different connotations. Going beyond the knowledge that is strictly connected to the cigar recommendations process and the panel, allow us to share a few words from a few of our members :
“Cigar Sense is a unique service that has helped me grow as a cigar smoker. I’ve discovered what flavors and […]. The best part of being a Cigar Sense member, beyond the great recommendations, is […] that Cigar Sense is not just some business looking to profit off the backs of cigar smokers. It is Franca Comparetto, who brings her knowledge, enthusiasm, and advocacy through Cigar Sense that makes it a personal experience for me and other cigar enthusiasts.” – Art Schlussel
“What I like the most about Cigar Sense is that it is helping me both learn and refine my taste preferences. Cigar Sense has really aided me in identifying key tastes and aromas […]” – Nikki Glenn
“[…] What you’ll learn will help you understand truly what you like, and will let you experiment with new tastes!” – Scott Krinsky
“[…] I have learned about myself, what I might like and have begun to get away from some of the usual’s.” – Christopher Lau
“[…] I also particularly enjoy the articles that are sent out via email. Those are interesting and help elevate my appreciation of why I enjoy cigars.” – Jonathan Faber
We are very gratified by all our members testimonials and feedback!