Framework

The Growth Loop: how e-commerce brands grow on every channel

By . Published June 28, 2026, rewritten October 8, 2026. Kestrel Kitchen figures render from the sample report data.

Part 1

Add up the whole business before you diagnose any channel

Do this first

  1. Export the last 13 weeks of sales from every channel you sell on.
  2. Add them into one weekly total, beside the same 13 weeks last year.
  3. If the total held, find where shoppers moved: a price gap, a one-channel promotion, or shifted ad budget.
  4. If the total fell, go to Part 2.
Whole business, last 13 weeks $160,318 +11.3% vs last year
Amazon alone $78,481 -20.8% vs last year
$0k $8k $16k Jun 7. Amazon falls and the storefront rises. Oct 24Jan 25Apr 25Jul 25Oct 25Jan 26Apr 26Jul 26
  • Amazon $78,481 -20.8%
  • Shopify storefront $54,709 +121.1%
  • Subscription renewals $23,431 +42.5%
  • Wholesale (Faire) $3,696 flat
Weekly revenue, 4-week rolling average, stacked by channel. The dark line is the whole business. Legend totals are the last 13 weeks against the same weeks a year earlier. Kestrel Kitchen is a fictional brand.
  • Each channel names the same numbers differently. Amazon says Unit Session Percentage, Shopify says conversion rate, Faire says reorders.
  • A falling channel can sit inside a growing business. Kestrel Kitchen, the fictional brand in our sample report, fell 20.8% on Amazon over 13 weeks against last year, while the whole business grew 11.3% to $160,318.
  • Shoppers who buy across channels spend more. Across 46,000 shoppers at one US retailer, they spent 4% more per store visit and 10% more online.1
  • One channel's data misreads a change. When a US home-goods retailer launched buy-online-pickup-in-store, online sales fell while store sales rose, and the researchers warned against "single-channel data."2

Part 2

Split sales into three numbers every channel already reports

Do this first

  1. Pull traffic, conversion and order value for the channel that fell, from the report named below.
  2. Compare each with the same weeks last year.
  3. Circle the one that fell. The levers under it are your suspects.
Sales
=
Traffic Did enough shoppers see it? VisibilityOperations
×
Conversion Did the ones who saw it buy? PricingAssortmentRatingsOperations
×
Order value How much did each order carry? PricingAssortment
The same three numbers, named the way each channel reports them
ChannelTrafficConversionOrder value
AmazonBusiness Reports, Detail Page Sales and Traffic SessionsUnit Session PercentageAverage selling price × units per order
ShopifyAnalytics, Reports SessionsConversion rateAverage order value
Wholesale (Faire)Faire brand portal, Orders Active retailersReorder rateAverage wholesale order
Retail and groceryRetailer portal or distributor report Stores carrying youUnits per store per weekShelf price
Every lever moves at least one of the three. Operations moves two at once, which is why a stockout looks like a traffic problem and a conversion problem on the same day.
  • Sales is traffic times conversion times order value on Amazon, on your own store, in wholesale and on a grocery shelf.
  • The split cuts 5 suspects to 1 or 2, because different levers move each number.

Part 3

Run five levers as one loop, in order

In stock and fairly priced, a product converts and earns rank. Found by more people, it collects reviews that say what to fix next. Find your first lever in 5 questions with the growth lever finder.

The Growth Loop Five levers arranged clockwise in a loop: Operations, Pricing, Assortment, Visibility and Ratings. Ratings feeds back into Operations. In the centre: one business, reviewed weekly. ONE BUSINESS reviewed weekly every channel, one set of numbers 1 Operations in stock, shipping fast 2 Pricing obviously worth it 3 Assortment the mix they want 4 Visibility found where they look 5 Ratings what they tell you
Five levers, one loop. Each feeds the next, and Ratings tells you which one is leaking. Select a lever to jump to it.

Part 3 · Lever 1

Operations: stay in stock, because every other lever stands on it

Do this first

  1. List your top 10 items by sales and their days of cover at current sell-through.
  2. Flag any item under 98% in stock over the last 30 days.
  3. Set each reorder point at what the item sells over your supplier lead time plus 2 weeks.
  • A stockout costs future orders. In a field test with 22,921 customers of a national mail-order catalog, a customer whose every ordered item was out of stock ordered about 22% less over the next 13 months.3
  • On Amazon it costs rank. Across more than 290,000 grocery products in 5 Amazon marketplaces, items unavailable for 90 days ranked 14% to 67% worse.4
  • Off Amazon it costs the shelf. 31% of shoppers who hit an empty shelf go to another store, and in wholesale it's a reorder that doesn't come.5

Go deeper: why you lost the Amazon Buy Box, which a stockout is one route to.

Part 3 · Lever 2

Pricing: be obviously worth it next to the alternative

Do this first

  1. Price your top 3 items, delivered, against the 2 closest alternatives your shoppers compare.
  2. Pull the share of orders carrying a discount code, by channel.
  3. Cut any standing discount that runs every week.
  • Price carries the most profit. At the average economics of 2,463 Compustat firms, McKinsey found a 1% price gain lifts operating profit 11.1%, against 3.3% for 1% more volume.6
  • A 1% price change moves units by more than 2%. A meta-analysis of 1,851 price elasticities put the average at -2.62.7
  • Promotions leave little behind. Their permanent effects on sales are "virtually absent,"8 and in 3 field experiments deeper discounts brought first-time customers back but cut future purchases by established ones.9
  • A gap between channels moves shoppers. In the sample, Kestrel puts a code on 22.47% of its Shopify orders and ran 0 promotions on Amazon in 12 months.

Go deeper: the five Buy Box checks, starting with landed price.

Part 3 · Lever 3

Assortment: carry the mix your customer came for

Do this first

  1. Find the 2 variants the category leader sells that you don't.
  2. Tally the complaints and questions that repeat in your reviews.
  3. Compare each bundle's conversion with the singles inside it.
  • The case for cutting SKUs reversed. An online grocer's SKU cut, first reported as an 11% sales gain, was corrected to about 1.5% (not significant), and a follow-up found sales fell and customers waited 25.0% longer between orders.10
  • Choice overload averages out near zero. The jam study (30% buying from 6 jams against 3% from 24) rests on 31 buyers against 4, and a meta-analysis of 50 experiments put the average effect at virtually zero.11
  • New products fail less than the folklore says. Studies since 1977 put the rate at 40% or less, against the 80% or higher that gets repeated.12 Launch against a repeated complaint or an uncovered search.
  • Bundles get the same test. In the sample, Kestrel's Care Kit converts at 3.93% on Amazon against 10.33% to 12.52% for the singles inside it, so the bundle is costing it shoppers.

Part 3 · Lever 4

Visibility: get found where your customer looks

Do this first

  1. Write down your top 3 search terms per channel and your rank on each.
  2. Split ad spend into brand and non-brand terms.
  3. Measure each channel's ad spend against that channel's margin.
  • Marketplace search rewards the other 4 levers. It ranks each item by how likely it is to satisfy the shopper who searched.
  • Paid search returned less than it cost in eBay's 2015 experiments. Brand-keyword ads had "no measurable short-term benefits," and non-brand ads lost money on average, because most spend reached frequent buyers who'd have bought anyway.13
  • Position moves clicks more than purchases. Randomized rankings on Expedia changed which hotels people clicked, with no effect on what they bought once they'd searched.14

Go deeper: why your Amazon sales are down, when the fork points at traffic.

Part 3 · Lever 5

Ratings: read what your customers say is broken

Do this first

  1. Read your last 50 reviews and tally the complaints.
  2. Fix the top complaint at the product, the listing or the packaging.
  3. Ask every buyer for a review through the channel's own tool, never with an incentive. The FTC's fake-review rule has applied since October 21, 2024.18
  • Each star is worth revenue. At Harvard, Michael Luca found each extra Yelp star lifted revenue 5% to 9% for independent restaurants, while chains barely moved.15
  • Bad reviews weigh more than good ones. One-star reviews hurt book sales more than five-star reviews helped them.16
  • The first 5 reviews count most. A product with 5 was 270% more likely to be bought than one with none, peaking at 4.2 to 4.7 stars (Northwestern's Spiegel Research Center, on PowerReviews data, not peer-reviewed).17
  • Kestrel's reviews already say why. In the sample, Board Cream fell from 4.6 to 4.2 stars on Amazon in 2026, on 1,034 reviews, while 0 of 4 competitors fell.

Go deeper: why Amazon reviews don't show up, what your next star is worth, and what 3,725 reviews reveal about listing accuracy.

Part 4

When sales drop, run the fork before you change anything

Do this first

  1. Answer the 3 questions below for the channel that fell.
  2. Open the lever it names, and leave the other 4 alone this week.
  3. On Amazon, run the Sales Drop Diagnostic on your Seller Central numbers for the revenue gap.
1 Which channel fell?
2 Across every channel combined, did the total business fall?
3 In that channel, compared with the same weeks last year, what fell?

Answer the questions and the matching lever appears here. Without JavaScript, all five answers are listed below.

  1. Total business held or grew

    Channel mix

    The total held, so shoppers moved between channels. Before you fix the falling channel, find where its customers went and why, using the same 13 weeks last year as the baseline.

    Check first

    • Price for the same item on each channel, this week
    • Which channel grew in the same week the other one broke
    • Discount codes, promotions or ad budget that started or moved that week

  2. Traffic fell, conversion held

    Visibility

    Fewer people are seeing you, and the ones who do still buy at last year's rate. The page still works, and the problem is getting found.

    In the sample, Kestrel's Amazon sessions fell 30.8% while conversion rose, which is this pattern.

    Check first

    Amazon

    • Organic rank on your top 3 search terms, week over week
    • Sponsored Products spend and impressions, by campaign
    • Any suppressed listing, lost badge or lost placement

    Shopify

    • Sessions by source in Shopify Analytics: Google, Meta, email, direct
    • Google Search Console clicks and impressions on your top pages
    • Paid social and email sends, week over week

    Wholesale

    • New retailers from Faire search, against new ones from your own outreach
    • Your Faire search placement and any lost badge
    • Retailers who stopped reordering, by name

    Retail

    • Stores dropped from the planogram or set
    • Out-of-store weeks at the distributor
    • Promotional slots or end caps you lost this period

  3. Conversion fell, traffic held

    Pricing, Assortment or Ratings

    The same people arrive and fewer of them buy. Something on the page or shelf stopped persuading. Check the fastest-moving cause first.

    Check first

    Amazon

    • Your price against the item next to you in search, today
    • Buy Box (Featured Offer) percentage by day
    • Your 10 newest reviews and whether the displayed rating moved

    Shopify

    • Your price against the same item on Amazon and at competitors
    • Checkout drop-off and shipping cost shown at checkout
    • Reviews and ratings shown on your top product pages

    Wholesale

    • Wholesale price and minimums against comparable brands on Faire
    • Retailer sell-through, since slow movers don't get reordered
    • Lines retailers asked for that you don't carry

    Retail

    • Shelf price against the brand next to you
    • Promotion calendar, yours and the competitor's
    • Facings and shelf position at your top stores

  4. Both fell at once

    Operations

    A stockout takes down traffic and conversion together, so start with inventory before you touch anything else.

    Check first

    Amazon

    • In-stock rate and any zero-inventory days on your top items
    • FBA inbound shipments stuck or delayed
    • Buy Box (Featured Offer) percentage by day

    Shopify

    • Sold-out variants on your top product pages
    • Delivery promise and fulfillment speed shown at checkout
    • Site errors or a broken checkout in the same week

    Wholesale

    • Backorders and unfilled Faire orders
    • Ship times against what retailers were promised
    • Items marked out of stock in your Faire catalog

    Retail

    • Distributor fill rate and short shipments
    • Out-of-stock stores in the retailer portal
    • Purchase orders the retailer cut or delayed

  5. Neither fell, so order value did

    Pricing or Assortment

    The same shoppers bought at the same rate and spent less each time. That's price, promotion depth or the mix in the basket.

    Check first

    • Discount rate and the share of orders using a code
    • Bundle and multi-pack share of units
    • Whether your top-priced item lost share to a cheaper one
Compare against the same weeks last year, not last month, so seasonality and event weeks don't pass for a problem.

Kestrel, run through the fork

  • Question 2: the whole business grew 11.3%, so Kestrel's first answer is channel mix.
  • Question 3 on Amazon: sessions fell 30.8% while Unit Session Percentage rose 129 bps to 12.41%. Fewer shoppers reached the listing, and more of them bought, which points at Visibility.
  • The export that settles it: ad spend by channel by month. Shopify orders rose 78.0% over the same 13 weeks, and budget moving to the storefront would explain both.

Part 5

Run it every week, because the returns compound

Do this first

  1. Put the 6 numbers in the table below on one page.
  2. Change one thing per lever per week.
  3. Read each change against the same weeks last year, and keep what moved the number.
  • Firms that decide from data produce more. Across 179 large US firms, they ran 5% to 6% above the output their other investments predicted.19
  • Structured reviews explain productivity. Across 35,000 US manufacturing plants, monitoring, targets and reviews explained more than 20% of the variation in productivity, as much as R&D or IT or more.20
  • Most changes won't work. At Google and Bing only 10% to 20% of experiments produce a positive result, and at Microsoft about a third do.21
A weekly loop against one-off tactics An illustration. A loop run every week climbs in big steps early, then compounds. One-off tactics spike and fade back to where they started each time. The loop, run weekly One-off tactics Weeks, then years
Illustrative, not brand data. Big steps early, while the obvious leaks close, then something closer to compound interest.

The weekly review, on one page

LeverThe number to readWhat good looks like
Whole businessNet revenue across every channel, against the same weeks last yearGrowing, with no channel over 70% of it
OperationsIn-stock rate, days of cover98%+ in stock, cover longer than your reorder lead time
PricingPrice against the comparable item, discount rateAt or under the comparable, or a reason not to be
AssortmentConversion by item, bundle shareNo item converting at under half of your best seller
VisibilitySessions by channel, organic rankPage one for your top 3 terms, paid share falling over time
RatingsStar rating by item, review velocity4.3 or higher, and a steady flow of new reviews

Directional, and they vary by category. Your own trend against last year matters more than any benchmark.

Part 6

Where this fits with the frameworks you know

  • The Amazon flywheel. Jeff Bezos's sketch describes a marketplace, where experience brings traffic, sellers and selection. The Growth Loop describes the 5 levers one brand controls.
  • The funnel. Traffic, conversion and order value are Part 2 of the loop, the part that says where to look.

We run this loop for you, every month

Send us Amazon's Sales and Traffic by Date report and your Shopify orders export. We add up the whole business, run the fork on every channel and hand back ranked actions, like the Kestrel sample. The first month is free, and we never log in to your accounts.

Get your first month free See the sample report

Sources

  1. Sopadjieva, E., Dholakia, U. M. and Benjamin, B. (2017). A Study of 46,000 Shoppers Shows That Omnichannel Retailing Works. Harvard Business Review. Link
  2. Gallino, S. and Moreno, A. (2014). Integration of Online and Offline Channels in Retail. Management Science 60(6). Link
  3. Anderson, E. T., Fitzsimons, G. J. and Simester, D. (2006). Measuring and Mitigating the Costs of Stockouts. Management Science 52(11). Link
  4. Etumnu, C., Chenarides, L. and Jaenicke, E. (2025). Grocery Stockouts and Seller Performance in Amazon's Marketplaces. Agribusiness. Link
  5. Gruen, T. W., Corsten, D. and Bharadwaj, S. (2002). Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes, and Consumer Responses. GMA, FMI and CIES. Offline grocery, 71,000 consumers in 29 countries. Link
  6. Marn, M. V. and Rosiello, R. L. (1992). Managing Price, Gaining Profit. Harvard Business Review. Link
  7. Bijmolt, T. H. A., van Heerde, H. J. and Pieters, R. G. M. (2005). New Empirical Generalizations on the Determinants of Price Elasticity. Journal of Marketing Research 42(2). Link
  8. Pauwels, K., Hanssens, D. M. and Siddarth, S. (2002). The Long-Term Effects of Price Promotions on Category Incidence, Brand Choice, and Purchase Quantity. Journal of Marketing Research 39(4). Link
  9. Anderson, E. T. and Simester, D. I. (2004). Long-Run Effects of Promotion Depth on New Versus Established Customers. Marketing Science 23(1). Link
  10. Borle, S., Boatwright, P., Kadane, J. B., Nunes, J. C. and Shmueli, G. (2005). The Effect of Product Assortment Changes on Customer Retention. Marketing Science 24(4). Reports the correction to Boatwright and Nunes (2001), Journal of Marketing 65(3), published by those authors as a 2004 correction note. Link
  11. Scheibehenne, B., Greifeneder, R. and Todd, P. M. (2010). Can There Ever Be Too Many Options? Journal of Consumer Research 37(3). The jam study is Iyengar, S. S. and Lepper, M. R. (2000), Journal of Personality and Social Psychology 79(6). Link
  12. Castellion, G. and Markham, S. K. (2013). New Product Failure Rates: Influence of Argumentum ad Populum and Self-Interest. Journal of Product Innovation Management 30(5). Link
  13. Blake, T., Nosko, C. and Tadelis, S. (2015). Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment. Econometrica 83(1). Link
  14. Ursu, R. M. (2018). The Power of Rankings: Quantifying the Effect of Rankings on Online Consumer Search and Purchase Decisions. Marketing Science 37(4). Link
  15. Luca, M. (2016). Reviews, Reputation, and Revenue: The Case of Yelp.com. Harvard Business School Working Paper 12-016. Link
  16. Chevalier, J. A. and Mayzlin, D. (2006). The Effect of Word of Mouth on Sales: Online Book Reviews. Journal of Marketing Research 43(3). Link
  17. Spiegel Research Center, Northwestern University Medill (2017). How Online Reviews Influence Sales. Built with PowerReviews data, not peer-reviewed. Link
  18. Federal Trade Commission (2024). Consumer Reviews and Testimonials Rule, 16 CFR Part 465. Effective October 21, 2024. Link
  19. Brynjolfsson, E., Hitt, L. M. and Kim, H. H. (2011). Strength in Numbers: How Does Data-Driven Decisionmaking Affect Firm Performance? MIT. Link
  20. Bloom, N., Brynjolfsson, E., Foster, L., Jarmin, R., Patnaik, M., Saporta-Eksten, I. and Van Reenen, J. (2019). What Drives Differences in Management Practices? American Economic Review 109(5). Link
  21. Kohavi, R. and Thomke, S. (2017). The Surprising Power of Online Experiments. Harvard Business Review. Link

Start here

Browse all guides and tools