science from sales and the microbiome
TRANSCRIPT
Science from sales and the microbiome
Leveraging the direct-to-consumer model to obtain actionable data
Justin Shmalberg | DVM DACVN DACVSMR CVA CVFT CVCHRyan W. Honaker | just PhD
Outline
1. NomNomNowOverview
2. Data collectionSign-up flow and health assessment
3. MicrobiomeCollection, studies, data
● Fresh made-to-order direct-to-consumer pet food
○ Weekly cooks in wholly-owned production facilities
○ Human restaurant and bulk foods suppliers
○ Innovative software engineering
● Customer base
○ Engaged
○ Demand Transparency
○ Seeking Customization
● Customization critical to expanding
reach to pet health from pet food
○ Product validation
○ Clinical research
Foundation for customization and studies
● Dedicated internal software engineering team
● Scientific team○ Veterinary nutritionist
○ Microbiologist
○ Bioinformaticist
○ Nutrition researcher
○ Advisory team
● Products guided by nutrition, data, and science to meet consumer needs
Customization: Calorie Use Case● Default calorie guidance unreliable
○ Breed variation
○ Activity variation
○ Life stage
● Pet parents unreliable
○ Confusion over feeding charts
○ Measure volumes incorrectly
○ Obesity!
Solution = Custom portioning
● Custom portions = custom algorithm = data driven
○ Age (life stage), activity, breed, gender, current weight, ideal
weight
○ Collection of useful data
What can more and better data achieve?● Optimize product success
○ Prevent incorrect use cases
■ Difficult in retail environment without human
support
○ Predictive factors for better outcome on certain diets
■ Body weight and fiber
■ Skin and coat and fatty acids
● Nutritional validation
○ Blood values?
■ Unlikely to change
○ Skin and coat?
■ May not be predictive of underlying health
○ Microbiome
■ Can this explain differential responses and possible benefits from diet?
Health assessments● Metadata for microbiome initiative
● 115 questions - 5 categories
● Species specific
● Results stored in SQL database
○ Results are cleaned with with Python
scripts
○ Easily queryable for analysis
Health assessments● Completion rate:
○ Overall: ⅔ none, 15% all
○ MB: ⅓ none, ¼ all
● Exploring ways to increase
○ Just ask: 12% complete additional
○ Offer incentives
○ Provide insights from results
● Provide customer and product insights○ Largest medical issue complaint is allergies (19%)○ 23% of customers are using probiotics
Utilization of health assessment results - product insights
Medical conditions
Supplements
● Generate content● Develop products
Utilization of health assessment results - customer facing
● Deliver personalized information based on responses○ Provide customer with value and purpose for taking surveys○ Cross selling products
● Identify needs and interests, e.g. GI issues and probiotics● Provide personalized: interesting data visualizations, content,
product recommendations
Microbiome Profiling - NomNomNow: INSIGHTS
● Consumer product:○ Kit○ Health assessment○ Microbiome report
NomNomNow: INSIGHTS report
A B C
D
A. Your dogB. All dogsC. All healthy dogsD. All 18 year-old 3-legged
neutered male shih tzus with IBS who chase their tails and eat sauerkraut
● Comparisons to various cohorts○ Composition○ Diversity○ Common health concerns
Benefits to customer:
● Comparison of their pet’s GI health
● Monitoring of MB over time
● Relationship to underlying conditions
● Citizen science and engagement
● Pet health customization - products
● As the science advances the results mean more and become increasingly actionable
NomNomNow: INSIGHTS
Benefits to us
● A rich source of data○ Scientific publication
● Funnel into food subscription
● Cross-selling products with impact
● Novel consumer engagement
● Product developments○ New diets, pro/pre/syn -biotics,
diagnostic markers
● Development of pet health brand
NomNomNow: INSIGHTS
Leveraging survey data for scientific study
1. Develop a testable hypothesis and define cohort
a. Retrospective/passive: The microbiome of dogs >30 kg differs from
that of dogs <10 kg
b. Prospective/targeted: Fresh diets alter the composition of the
microbiome
c. Discovery: machine-learning predictions for novel associations
Survey response: interested in studies
4. Reach out to customers
○ Personalized emails (customer/pet
name)
○ Custom copy relevant to study
○ Incentivize (free sequencing, discounted
food/treats, etc.)
Leveraging survey data for scientific study
Recruitment success rate: ⅓ of cats, ½ of dogs
Model AIC = 1044.2
Adjusted R squared = .21
p(age) = 2e-16***
p(size) = 2e-16***
p(periodontal) = 0.01*
~60%
~40%
~60%
~30%
Survey data insights - science, product development
Increased resolution improves results
Shotgunmeta-
genomics
individual pets
ind
ivid
ual
tax
a
imore blue = more bugs
Discovery - what factors are driving clustering?
Dogs
Clusters 1-2
Cluster 3
Green = high amount
Red = low amount
Black = absent
individual pets
ind
ivid
ual
tax
a
Goals and milestones - short term
● Scientific validation of claims and of efficacy
○ Education and outreach mission
○ Advance the science of fresh
● Report updates based on novel observations and areas
of customer interest
Goals and milestones - short term
● Identify and target unmet consumer needs: targets
for studies and product development
○ Therapeutic diets
○ Diagnostic markers
○ Therapeutic interventions
■ Pro/pro/synbiotic
■ Functional genes / products (postbiotics)
○ Intellectual property generation
● Building pet health and science brand
Goals and milestones - longer term
● Contribute valuable scientific information
○ How does the MB impact health and disease?
○ Role of nutrition and the MB
○ Human and animal translational research?
Goals and milestones - longer term
● Develop products for unmet consumer needs:
○ Therapeutic diets
○ Diagnostic markers
○ Therapeutic interventions
■ Pro/pro/synbiotic
■ Functional genes / products (postbiotics)
○ Personalized nutrition / supplements -
the ultimate in customization
Lessons learned● Direct to consumer sales platform invaluable in truly knowing your customer
● Pet owners readily provide detailed pet information in a user-friendly interface --
At least ⅓ of engaged customers interested in scientific studies
● New methods of assessing nutritional performance are needed --
The microbiome knowledge of pet dogs and cats remains in its infancy
● Big data and meaningful science doesn’t have to mean big budgets
● Producers of pet products have a responsibility to help advance the science
○ Especially when validating new product forms or marketing claims
Acknowledgements
R&D CX Ops, marketing
Aashish Jha Melanie De Anda Trevor Hargreaves
Dan Massey Jessica Siasoco Ryan Bauer
LeeAnn Perry
Jirayu “Boo Boo” Tanprasertsuk