Evergreen
How to Compare Grocery Prices Across Stores
Compare a real grocery basket across stores by normalizing sizes, fees, travel, substitutions, memberships, availability, and price freshness.
How to Compare Grocery Prices Across Stores
To compare grocery prices across stores, price the basket you will actually buy, normalize package sizes, collect the offers at the same time, and add membership costs, fees, tips, travel, missing items, substitutions, and uncertainty. The cheapest displayed basket is not necessarily the cheapest practical trip.
The goal is not to monitor every price. It is to build one repeatable comparison that reflects your household and can be refreshed when stores, fees, or habits change.
Start with a recurring basket, not sale items
A comparison becomes useful when the items recur. Choose groceries that represent an ordinary week or month. Include products where the exact brand matters, products where any acceptable equivalent will work, and at least one item that is frequently out of stock.
Do not build the basket from whatever is promoted today. A promotion-led list measures the retailer's current marketing, not your normal spending. It also makes the next comparison hard to repeat.
Record the quantity and the acceptable substitution rule for each item. “Milk” is too vague if one store returns a half gallon and another returns a gallon. “Cereal” is too vague if one household will accept any variety and another needs a specific allergen profile.
Unit price makes unlike packages comparable
The National Institute of Standards and Technology describes unit pricing as the cost per standard measure, such as per ounce, pound, liter, count, or square foot. It helps shoppers compare differently sized packages and notice shrinkflation.
Use the same unit within an item category. Divide the package price by net quantity when the retailer does not display a usable unit price. Check that the package is truly comparable. A larger package may have a lower price per ounce but create waste, storage pressure, or a higher checkout total.
For food that changes substantially during preparation, cost per serving may be more useful than package weight alone. The USDA Food Buying Guide includes yield and cost-per-serving concepts for food-service planning. A household does not need its full institutional method, but the principle is valuable: raw purchase quantity is not always edible or usable quantity.
Capture comparable offers
The price record should identify the exact product, package, store location, fulfillment channel, account state, membership, promotion, coupon, and time. Comparing an in-store shelf price with a delivery-marketplace price may reveal a real household choice, but it does not isolate the retailer's item price.
Collect prices in one bounded session. If the comparison takes several days, fast-moving promotions and inventory updates can make the result internally inconsistent. Save screenshots or notes for surprising differences.
Retail platforms can show different offers under different conditions. The 2025 Consumer Reports Instacart investigation found simultaneous item-price variation during coordinated shopping sessions. Instacart said the differences came from randomized tests rather than personal-data pricing. The dispute is a reminder that a displayed price needs a timestamp and context.
flowchart LR
A["Recurring household basket"] --> B["Exact item and unit"]
B --> C["Same-time store offers"]
C --> D["Memberships, coupons, and fees"]
D --> E["Travel, time, and accessibility"]
E --> F["Missing items and substitutions"]
F --> G["Confidence-adjusted trip decision"]
The comparison becomes more realistic at each step. It also becomes more personal.
Use a seven-step comparison
- Define the recurring basket.
Write the product, quantity, acceptable brands, required dietary attributes, package range, and substitution rule. Mark an item “exact” when a substitute would make the comparison invalid.
- Normalize quantity and quality.
Calculate a common unit price. Check net weight, count, formulation, freshness, and package type. Do not treat two items as equivalent because their names are similar.
- Capture prices in the same time window.
Use the same store location and fulfillment channel for each option. Record whether you are signed in, whether a loyalty card is active, and whether the price depends on a coupon or quantity.
- Add transaction costs.
Include delivery, service, small-order, bag, pickup, and membership costs that apply to this basket. Include an expected tip when delivery is the real option. Taxes vary by product and jurisdiction, so use the amount that will actually be charged rather than a universal assumption.
- Add trip costs and constraints.
Record walking, driving, parking, tolls, transit, or rideshare. Time should be treated as a household input, not assigned one universal wage. Mobility, childcare, caregiving, weather, safety, and schedule can determine whether an option is feasible.
- Apply missing-item and substitution rules.
A store that lacks two essential products may force a second trip. Price that outcome. If a replacement is acceptable, use its real unit price and note the quality difference. If availability is uncertain, do not give the total the same confidence as a fully confirmed basket.
- Choose the option and preserve the record.
Select the store or channel that fits the whole decision. Save the basket, date, and assumptions so the next check updates a known baseline rather than starting over.
A compact worksheet prevents false precision
| Comparison field | Store A | Store B | Delivery option |
|---|---|---|---|
| Comparable item total | |||
| Membership and coupon effect | |||
| Service, delivery, and other fees | |||
| Expected tip | |||
| Travel or access cost | |||
| Missing-item or second-trip cost | |||
| Total practical cost | |||
| Price and stock confidence |
The worksheet does not need to monetize every inconvenience. A plain-language note such as “not feasible without a car” can be more honest than assigning a speculative dollar amount.
Distance can outweigh the displayed savings
In episode 99, Dalton describes comparing Stretch results in New York City without a car. A store with a lower basket could still be the worse choice if it required a longer trip. Andy Ellwood calls this “proximity pricing preference.”
The idea applies beyond Manhattan. A rural shopper may face a long drive. A person with limited mobility may value a reliable delivery window. A caregiver may need the nearest store that consistently has the exact formula or dietary product. A household making one monthly stock-up may accept a trip that makes no sense for a small weekly basket.
Stretch's current website says the app tracks local stores and shows prices and locations. Its terms also say retailer and manufacturer data can contain errors or expired prices and instruct users to verify an offer before purchase. Any comparison tool should expose that freshness and coverage boundary.
Treat membership and coupons as conditional prices
A loyalty price is not the base price. It is an offer available under stated conditions. Record the membership cost, renewal period, minimum spend, redemption limit, and whether the benefit is already sunk for your household.
A coupon can require clipping, a linked account, a minimum quantity, a specific pickup window, or the sharing of purchase data. The FTC's online shopping guidance recommends comparing policies and total terms, not only the headline price. The same discipline belongs in grocery comparison.
Do not count a first-order promotion as a recurring saving. Do not annualize one unusually favorable trip. If a platform changes its terms or fees, update the comparison rather than preserving a stale winner.
Uncertainty deserves its own column
A precise total can be wrong. Give each store a confidence note based on price timestamp, stock visibility, substitution risk, and how often the observed total survives checkout.
When two choices are close, reliability may be the deciding factor. A basket that is two dollars cheaper but frequently loses three exact items is not clearly cheaper. A price comparison product should show which products matched exactly, which were inferred, which prices are promotional, and when the data was last checked.
This is also where household correction matters. If an app learns that you usually buy one brand but the preference has changed, the comparison should be easy to repair. [[What a Trustworthy Digital Pantry Requires]] explains the data and control system behind that experience.
Refresh the method, not every price
Run the full comparison monthly, quarterly, or after a material change. A new store, delivery fee, membership, dietary need, move, vehicle, or work schedule can justify a refresh. Between full checks, watch the few products and fees that drive the decision.
The comparison is successful when it reduces uncertainty and supports an actual trip. It is not successful merely because it contains more prices.
Episode 99's [[Andy Ellwood on Building Stretch After Basket|conversation with Andy Ellwood]] explains why grocery transparency remained interesting after Basket and why Stretch begins with the household's list. [[Algorithmic Grocery Pricing - What Shoppers Can and Cannot See]] is the next read when the question is not which option costs less, but why two shoppers may see different prices.
Sources and editorial notes
This guide uses NIST unit-pricing guidance, the USDA Food Buying Guide, FTC online-shopping guidance, the Consumer Reports pricing investigation, Instacart's response, Stretch's current product materials, and the preserved episode 99 transcript.
It does not rank retailers, promise savings, assign a universal value to time, or provide individualized financial advice. Fees, prices, laws, and product capabilities require refresh before publication. AI assisted with research organization and drafting under editorial review. Publication remains unauthorized.
Sources
Follow the evidence.
- ag.ny.gov: attorney general james demands answers instacart about algorithmic pricingag.ny.gov
- nist.gov: nist sp 1181 unit pricing guide best practice approach unit pricing 2015 ednist.gov
- gs1.org: gtings1.org
- search.ftc.gov: instacart pay 60 million consumer refunds settle ftc lawsuit over allegations it engaged deceptivesearch.ftc.gov
- stretchgroceries.com: privacystretchgroceries.com
- stretchgroceries.comstretchgroceries.com
- linkedin.com: andyellwoodlinkedin.com
- consumer.ftc.gov: online shoppingconsumer.ftc.gov
- steveblank.com: customer development is not a focus groupsteveblank.com
- steveblank.com: customer discovery in the time of the covid 19 virussteveblank.com
- stretchgroceries.com: termsstretchgroceries.com
- andyellwood.comandyellwood.com
- consumerreports.org: instacart stops ai pricing experiments a1176475852consumerreports.org
- foundersfund.com: choose good questsfoundersfund.com
- nist.gov: uniform unit pricing tools consumers fight shrinkflationnist.gov
- gs1.org: global data modelgs1.org
- gs1.org: current standardgs1.org
- ftc.gov: surveillance pricingftc.gov
- nber.org: w24489nber.org
- nist.gov: best practices uniform unit pricing update nist sp 1181 and nist handbooknist.gov
- foodbuyingguide.fns.usda.gov: Aboutfoodbuyingguide.fns.usda.gov
- andyellwood.com: founderandyellwood.com
- company.instacart.com: the truth about pricing tests on instacartcompany.instacart.com
- ftc.gov: ftc surveillance pricing study indicates wide range personal data used set individualized consumerftc.gov
- apps.apple.com: id6748532450apps.apple.com
- consumerreports.org: instacart ai pricing experiment inflating grocery bills a1142182490consumerreports.org