Volume Growth

Is the company selling more units, serving more orders, or handling more activity? Volume growth strips out price changes, so it shows whether real demand is rising or whether higher prices are doing all the work.

The iPhone count that disappeared

On November 1, 2018, Apple reported a record year. In the fiscal year ending September 2018, iPhone revenue was about $166.7 billion, up 18%. But the same table told a quieter story. Apple sold 217.7 million iPhones, almost exactly the 216.8 million of the year before. Its annual report said iPhone sales rose "due primarily to a different mix of iPhones resulting in higher average selling prices." Average revenue per iPhone had jumped from about $652 to about $766.[^1] Unit sales were still below their fiscal 2015 peak of 231.2 million.[^2]

The pattern had been building for three years. Lay Apple's own numbers side by side and divide revenue by units:

Fiscal year	iPhones sold	iPhone revenue	Revenue per iPhone
2014	169.2 million	$102.0 billion	about $603
2015	231.2 million	$155.0 billion	about $671
2016	211.9 million	$136.7 billion	about $645
2017	216.8 million	$141.3 billion	about $652
2018	217.7 million	$166.7 billion	about $766

Sources: Apple Form 10-K filings for fiscal 2016 and 2018.[^2][^1] Revenue per iPhone is calculated.

Fiscal 2015 was a giant year, with units up 37%. Then units fell 8% in fiscal 2016. Apple's annual report blamed "a lower rate of iPhone upgrades" and tough economies in several big markets.[^2] Units barely moved after that. What changed was price. In September 2017, Apple launched the iPhone X "starting at $999," far above earlier flagship phones.[^3] Buyers who chose it lifted the average price by about $114 per phone in a single year.

That same November day in 2018, Apple said it would stop reporting how many iPhones, iPads, and Macs it sold. Chief financial officer Luca Maestri explained that "a unit of sale is less relevant today than it was in our past."[^4] Many investors suspected Apple was hiding a weak number. Together with a soft holiday forecast, the news sent the stock down almost 7% the next day.[^5]

Two months later, Tim Cook cut Apple's holiday-quarter revenue forecast to about $84 billion. Most of the shortfall was in Greater China, and Cook also blamed "fewer iPhone upgrades than we had anticipated."[^6] The stock fell about 10% in a day.[^7] In fiscal 2019, iPhone revenue fell 14%, "due primarily to lower iPhone unit sales."[^8]

The warning had been in plain sight. For three years Apple sold about the same number of phones, and higher prices did nearly all the growing. Prices can only rise so far before buyers wait longer to upgrade.

The story did not end there, and that matters too. Flat units told investors that the old formula, more phones at higher prices, had hit a ceiling. They did not say Apple was finished. In fiscal 2021, a new cycle of phones pushed iPhone revenue up 39% to a record $192.0 billion, and Apple's services business (the App Store, iCloud, music, and other subscriptions sold to people who already own its devices) reached $68.4 billion.[^9] Volume is a question, not a verdict. It tells you what kind of growth you are looking at, and therefore what could go wrong next.

What volume is

Volume is the amount of activity a business handles, counted in units rather than money: phones shipped, coffees sold, passengers flown, rail cars moved, payments processed, or hours of software use.

Revenue mixes two things together: how much a company sells and how much it charges. Volume isolates the first. Economists make the same split for a whole country, separating "nominal" growth, measured in money, from "real" growth, which removes price changes and counts actual goods and services. A company's volume growth is its own "real" growth.

A good volume measure has three features. It counts something customers choose to do, such as visiting a store or shipping a box. It is defined the same way every period. And it is close to what drives costs and revenue, so a change in the count means a change in the business. "Number of stores" fails the first test, because a company can open stores whether or not customers show up. "Customer visits per store" passes all three.

How to measure it

Volume growth is the percentage change in units from one period to the next:

Volume growth = (This period's units − Last period's units) ÷ Last period's units

It links to revenue through a simple identity:

Revenue = Volume × Average price per unit

So revenue growth is roughly volume growth plus price growth. The exact version multiplies instead of adds:

(1 + Revenue growth) = (1 + Volume growth) × (1 + Price growth)

Apple's fiscal 2018 shows how it works. Units rose about 0.4% and revenue per phone rose about 17.4%. Multiply 1.004 by 1.174 and you get about 1.18, the 18% revenue growth Apple reported.[^1] Almost all of it was price.

Imagine a pizza shop that sold 10,000 pizzas at $15 last year, for $150,000. This year revenue is $165,000, up 10%. It could get there three ways:

• Volume-led: 11,000 pizzas at $15. Volume +10%.
• Price-led: 10,000 pizzas at $16.50. Volume flat.
• Price masking a decline: 8,800 pizzas at $18.75. Volume −12%, price +25%.

The headline is the same 10%, but in the third case 1,200 fewer pizzas left the oven.

Two more tools help. The first splits volume for a chain into locations and activity per location:

Volume = Number of locations × Units per location

If the pizza company opens a second shop and total pizzas rise 10%, but each shop sells fewer pizzas than the old shop did alone, demand per shop is falling.

The second is the two-year stack, which links two years of growth to see through an easy or hard comparison:

Two-year growth = (1 + Last year's growth) × (1 + This year's growth) − 1

If the shop's pizzas fell 10% last year and rose 10% this year, the stack is 0.90 × 1.10 − 1 = −1%. The shop is still slightly below where it started, even though this year's number looks strong.

Each industry has its own measure. Restaurants report comparable transactions (visits at stores open at least a year), railroads carloads, software companies active users, and aircraft makers deliveries (see Backlog Conversion). Use the same measure, defined the same way, for at least three periods.

Why volume growth matters

Companies set prices, but customers decide volume. That makes volume an early, honest signal of demand.

Professional investors pay for that signal. Researchers studied 4.8 million satellite images of parking lots at about 67,000 stores of 44 major U.S. retailers from 2011 to 2017. Year-over-year changes in car counts, a pure volume measure, reliably predicted quarterly sales. Investors with the data could earn about 4% to 5% in the three days around earnings announcements, especially by betting against retailers about to report bad news.[^10]

Volume also improves forecasts. Asher Curtis, Russell Lundholm, and Sarah McVay built a retail sales model that splits growth into more stores and more sales per existing store. Using only a few years of public data, it produced in-sample errors under 2% of sales, and combined with analysts' forecasts it was significantly more accurate than the analysts alone.[^11]

The same lesson shows up across industries. In a classic study of independent cellular phone companies, Eli Amir and Baruch Lev found that, on their own, earnings, book values, and cash flows were "largely irrelevant" for explaining share prices. Two non-money measures did the work: the population in each company's licensed area, a proxy for future growth, and market penetration, the share of that population actually signed up. Combined with those counts, earnings became useful again.[^12] For young companies whose profits have not arrived, counts of activity are often the best evidence available.

Airlines tell a similar story. Bruce Behn and Richard Riley tested the measures airline analysts watch. They found that load factor (the share of seats filled), market share, and capacity (available ton-miles) were tied to operating income and revenue in the same quarter. Using just one or two months of these operating figures, available well before the quarter's financial results, helped predict quarterly revenue, expenses, and operating income.[^13]

Volume data is only as good as its accuracy, though. Brett Trueman, Franco Wong, and Xiao-Jun Zhang studied internet companies from 1998 to 2000. Estimates of web traffic growth improved revenue forecasts over a model based on past revenue alone. But the traffic estimates then available generally added nothing beyond analysts' forecasts. Perfect knowledge of actual traffic growth would have helped, so the authors called for better and more timely usage reports.[^14] The lesson for investors: an exact count from the company beats a rough outside guess.

Why, then, do companies lean on price? Because price is powerful if volume holds. A McKinsey analysis of large U.S. companies found that a 1% price rise, with volume unchanged, would lift operating profit about 8%, more than three times the effect of 1% more volume.[^15] But volume rarely stays unchanged. A review of 1,851 estimates from 81 studies found an average price elasticity (how much buying falls when price rises) of about −2.6: a 1% price increase cut the amount of a brand people bought by about 2.6%.[^16] Volume is where the bill for overpricing shows up.

Volume bought with discounts is not free either. Harald van Heerde, Sachin Gupta, and Dick Wittink studied the jump in a brand's unit sales during a price promotion. Measured in units, about 33% of the extra sales came out of other brands in the same category.[^17] The rest came from shoppers buying more of the category or buying sooner, and buying sooner means part of the bump is borrowed from later weeks.

So volume growth is not automatically good. Discounts can buy volume that loses money on every sale, and some of it would have arrived anyway.

The three questions

1. Direction: is activity speeding up, steady, or slowing?

Look at the trend over at least three periods. Improving means underlying units, usage, or transactions are accelerating and spreading to more parts of the business. Stable means activity grows near its recent pace without a big change in what is sold. Deteriorating means activity is slowing or shrinking, even if price hides it.

Watch the base. Growth of 4% after a year of −4% only gets you back to where you started. Investors often combine two years, a "two-year stack," to see through easy comparisons. Watch the calendar as well. Some companies use a 52-week year that occasionally has a 53rd week, and that extra week can add one or two percentage points of growth that disappear the next year.

2. Breadth: is the growth widespread?

Volume growth is more convincing when it appears in existing and new stores, in several regions, and across loyal and occasional customers. If all of it comes from new stores while older ones lose visitors, the company is spreading thinner, not getting more popular.

Also check for a mix shift: a carmaker selling more cheap compacts and fewer trucks can show rising units and falling revenue. The reverse is common too. Apple in fiscal 2018 sold the same number of phones, but a richer mix of them.

3. Honesty: is price hiding the trend, or buying it?

Compare volume with price:

• Price masking a decline. Revenue rises while units fall, as at Apple. Each price increase may push more customers away.
• Discounts buying volume. Units rise only because of price cuts or promotions, and may vanish when they stop.
• Healthy growth. Units rise at full price.

Then compare with the industry cycle and peers. If a whole industry is losing traffic, a company that is merely flat may be winning share. If the whole industry is booming, a company growing slower than its rivals is losing ground even while its own numbers rise. The chapter on End-Market Demand covers how to read the industry backdrop, and Market Share covers how to tell share gains from a rising tide.

Taking volume apart

More locations versus more per location. Chains grow volume by opening locations or by serving more customers at each one. In the second quarter of 2026, Chipotle's revenue grew 9.3%, but comparable restaurant sales rose only 2.2%: 1.2% from a higher average check and 1.0% from more transactions. Most growth came from new restaurants; it opened 100 that quarter.[^18] New stores are real growth, but they cost money to build. Growth at existing stores is cheaper and a better sign of rising demand.

Capacity utilization. This is the share of possible output a business actually uses, like an oven that could bake 12,000 pizzas a year but bakes 11,000. Airlines call it the load factor, the share of seats filled. When utilization is low, extra volume is cheap to serve and spreads fixed costs over more units, an effect called Operating Leverage. Near full capacity, more volume requires new factories, planes, or stores.

Volume by industry. The right unit differs a great deal from one business to the next, and so do the traps:

Industry	Common volume measure	What to watch
Restaurants and retail	Comparable transactions or traffic; units per store	New stores can hide falling visits per store
Packaged food and drinks	"Organic volume" or unit cases	Price increases often hide shrinking volume
Railroads and trucking	Carloads, containers, ton-miles	Mix: a carload of coal earns far less than a carload of cars
Airlines	Revenue passenger miles (one paying passenger flown one mile); load factor	Seats added faster than passengers lower the load factor
Phones, PCs, chips	Units shipped	A richer mix lifts revenue with flat units
Software and internet	Active users, transactions, usage	Definitions of "active" can change
Oil and gas	Barrels of oil equivalent produced per day	Output can rise while oil prices fall

Packaged food shows the price-masking pattern clearly. In 2023, revenue at PepsiCo's Frito-Lay North America snack business grew 7%, "primarily driven by effective net pricing," while unit volume fell 1%.[^19] In 2024 the bill arrived. Frito-Lay North America's revenue fell 1% as unit volume dropped 2.5%, with declines in Cheetos, Tostitos, and Lay's, and its operating profit fell 7%.[^20] Two years of price-led growth had been hiding customers who bought fewer bags, and the hiding stopped working once pricing slowed. Price and Mix Growth follows the rest of that story, including the price cuts PepsiCo made to win the volume back.

Structural changes. Sometimes revenue and volume part ways because the business model changed, not demand. When a company sells its stores to franchisees or licensees, it stops booking those stores' full sales and collects fees instead. Revenue can fall even while customers visit more. The Starbucks example below includes such a change. In these cases look for system-wide sales (sales at all stores, owned or not) or per-store visits, not total revenue.

When falling volume did not mean a falling business: Union Pacific

Volume is a question, not a verdict, and Union Pacific, a major U.S. freight railroad, shows why. A carload is one rail car of freight, or one container on a train.

Year	Carloads (thousands)	Freight revenue	Revenue per carload	Net income	Operating ratio
2014	9,625	$22.6 billion	about $2,344	$5.2 billion	63.5%
2019	8,346	$20.2 billion	about $2,425	$5.9 billion	60.6%
2022	8,169	$23.2 billion	about $2,835	$7.0 billion	60.1%
2025	8,447	$23.2 billion	about $2,749	$7.1 billion	59.8%

Sources: Union Pacific Form 10-K filings.[^21][^22][^23][^24] Revenue per carload is calculated. The operating ratio is operating expenses divided by revenue, so lower is better.

Carloads peaked in 2014 and then fell. Coal was one big reason: in 2015 shipments from the Powder River Basin coal region fell 15%, which the company tied to low natural gas prices and high coal stockpiles.[^25] By 2019, carloads were 13% below 2014. A reader using this chapter's warning signs would see deteriorating volume.

Yet profit rose. In October 2018 Union Pacific began a cost program called Unified Plan 2020, built on precision scheduled railroading, which the company described as emphasizing on-time service for every rail car, accountability, and "lean resource utilization": doing the same work with fewer resources.[^26] In 2019 it reported earnings per share up 6% "despite volume declines of 6 percent," and a record operating ratio of 60.6%. Average employees fell from 47,200 in 2014 to 37,500 in 2019.[^21][^22] Revenue per carload kept rising, so the railroad earned more on each car it did haul.

So falling volume was not a death sentence. But it was not harmless either. A railroad run very lean has less slack when freight returns. In 2022, as demand recovered, Union Pacific's network clogged. Its operating ratio worsened 2.9 points, which it blamed partly on "operational inefficiency."[^23] To relieve congestion it turned to embargoes, temporary refusals to accept certain shipments. The federal rail regulator, the Surface Transportation Board, noted that Union Pacific's embargoes had risen "from a total of 5 in 2017 to more than 1,000 to date in 2022" and called a public hearing.[^27] In 2024 it said Union Pacific's embargoes in 2023 still outnumbered those of all other large railroads combined.[^28] By 2025, carloads were still about 12% below 2014.

The lesson cuts both ways. Shrinking volume can coexist with rising profit when a company has pricing power and room to cut costs, a theme of Pricing Power. But profit growth that relies on price and cost cuts while volume shrinks has limits, and the volume number tells you how much of the story rests on them.

Related theses. Volume is one half of Revenue Growth; the other is Price and Mix Growth. Volume from the existing business, excluding acquisitions, feeds Organic Growth. For subscription businesses, volume splits into Customer Base Growth and Existing-Customer Expansion. Because extra volume spreads fixed costs, it also feeds Operating Margin.

A worked example: Starbucks

Starbucks reports exactly this split. For U.S. company-operated stores open at least 13 months, it divides comparable sales into comparable transactions (volume) and average ticket (spending per visit). Its fiscal year ends in late September or early October.

Period	U.S. comparable sales	Comparable transactions	Average ticket
Fiscal 2018 (Americas)	+2%	−1%	+3%
Fiscal 2019	+5%	+2%	+3%
Fiscal 2020	−12%	−21%	+11%
Fiscal 2021	+21%	+8%	+13%
Fiscal 2022	+12%	+4%	+8%
Fiscal 2023	+9%	+3%	+6%
Fiscal 2024	−2%	−5%	+4%
Fiscal 2025	−2%	−4%	+2%
Q4 fiscal 2025 (Jul–Sep 2025)	0%	−1%	+1%
Q1 fiscal 2026 (Oct–Dec 2025)	+4%	+3%	+1%
Q2 fiscal 2026 (Jan–Mar 2026)	+7.1%	+4.3%	+2.7%
Q3 fiscal 2026 (Apr–Jun 2026)	+7.9%	+4.2%	+3.6%

Sources: Starbucks earnings releases.[^29][^30][^31] Fiscal 2018 is the Americas segment, the closest split Starbucks gave that year. Fiscal 2023 and 2024 figures cover North America and the U.S., which Starbucks reported together. Fiscal 2021 had 53 weeks.

Before the slump. Even in fiscal 2018, ticket did the work: Americas transactions fell 1% while ticket rose 3%. Fiscal 2019 was healthier, with transactions up 2%.[^29] Then the pandemic broke every comparison. U.S. transactions fell 21% in fiscal 2020 and rose 8% in fiscal 2021. Starbucks itself used a two-year view to make sense of it, headlining its fourth quarter of fiscal 2021 as "U.S. Up 22% with 11% Two-Year Growth."[^29]

The long view. Chaining the yearly figures gives a rough picture of an average comparable U.S. store from fiscal 2019 to fiscal 2025:

Measure	Calculation	Fiscal 2025 vs fiscal 2019
Transactions	0.79 × 1.08 × 1.04 × 1.03 × 0.95 × 0.96	about −17%
Average ticket	1.11 × 1.13 × 1.08 × 1.06 × 1.04 × 1.02	about +52%
Comparable sales	0.88 × 1.21 × 1.12 × 1.09 × 0.98 × 0.98	about +25%

This is rough arithmetic, not a reported figure. The set of comparable stores changes every year, the published numbers are rounded, and fiscal 2021 had an extra week. Still, the direction is unmistakable: sales per store rose by about a quarter while visits fell by about a sixth. Some of that is healthy, since a customer who visits less often may buy more per trip. But it is the Apple pattern again, in coffee instead of phones.

The deteriorating years. In fiscal 2023, two-thirds of comparable growth came from ticket. In fiscal 2024, revenue still rose 1% to $36.2 billion, but U.S. transactions fell 5%, and 10% in the fourth quarter. Starbucks ran more promotions, yet visit frequency "declined across customer segments," and it said its "traffic focused investments" further pressured results.[^32] Price was masking a decline, and discounts failed to buy volume back. In August 2024 the board hired Chipotle's chief executive, Brian Niccol, and the stock jumped more than 20% in a day.[^33]

Direction. U.S. transactions were still down 4% in fiscal 2025. The fourth quarter of fiscal 2025 was nearly flat, down 1%.[^31] Then they rose 3% in the first quarter of fiscal 2026, the first U.S. transaction growth in eight quarters,[^34] and about 4% in each of the next two quarters. The direction is improving.

But check the base. In the third quarter, transactions rose 4.2% after falling 3.7% a year earlier, a two-year stack of only about +0.3% (1.042 × 0.963 ≈ 1.003). Two years of losses (0.95 × 0.96 ≈ 0.91) followed by about 4% growth leaves traffic per store roughly 5% below fiscal 2023. That is a rough estimate, since the set of comparable stores changes each year. The recovery is real but unfinished.

Breadth. In the third quarter, transactions rose 4.5% in North America and 2.6% internationally. Starbucks had fewer U.S. stores than a year earlier (16,933 versus 17,230), yet U.S. revenue grew 7%, so more visits per store drove growth.[^35] China, where Starbucks had been adding stores fastest, had been a weak spot: its comparable transactions fell 22% in fiscal 2022, amid COVID-19 restrictions, and rose a modest 2.1% in the second quarter of fiscal 2026.[^29][^36] Peers were weaker. McDonald's U.S. comparable sales rose only 0.8% with "negative comparable guest counts," and Chipotle's transactions rose 1.0%.[^18][^37] That points to company-specific improvement, not just a better economy.

A structural change to adjust for. In the third quarter of fiscal 2026, Starbucks moved its China stores to a licensed joint venture model. That shift lowered company-operated store revenue, and consolidated revenue fell 1% to $9.3 billion even as U.S. transactions grew more than 4%.[^35] Company-operated stores fell to 33% of the global total, from 52% a quarter earlier.[^35][^36] A reader looking only at total revenue would see shrinkage. The per-store volume measures show the opposite, which is why this chapter leans on them.

Honesty. In each quarter of fiscal 2026, transactions grew faster than ticket. But ticket is climbing, up 3.6% in the third quarter, which Starbucks credited to delivery, food add-ons, and drink customizations.[^35] If ticket keeps rising while transactions fade, the fiscal 2023 pattern could return.

The cost of volume. Winning customers back was expensive. Starbucks invested heavily in store labor, and its non-GAAP operating margin (an adjusted figure that excludes one-time items) fell from 15.0% in fiscal 2024 to 9.9% in fiscal 2025.[^32][^31] In the third quarter of fiscal 2026, margin finally expanded, by 4.3 points to 14.4% on a non-GAAP basis, which Starbucks credited to "sales leverage": more visits spread over the same fixed costs.[^35] Its fiscal 2028 plan targets U.S. comparable sales growth of 3% or more and a 13.5%–15% non-GAAP operating margin.[^38]

The thesis. As of the third quarter of fiscal 2026 (the quarter ended June 28, 2026), an investor could state: "Starbucks' U.S. traffic recovery is durable, and customers keep visiting more often after the easy comparisons end." The resolution rule: "For fiscal 2027 (ending in late September 2027), U.S. comparable transactions rise at least 2% for the full year, with no quarter below zero." That is a real bar, since fiscal 2027 laps a strong fiscal 2026. It fits the typical 18-month horizon for a volume thesis.

The rule produces three clear outcomes. If transactions grow 2% or more with no negative quarter, the thesis is confirmed: customers are coming back more often, not just bouncing off a low base. If transactions grow but by less than 2%, or one quarter dips below zero, the result is mixed, and the next step is to check whether peers slowed too. If full-year transactions fall, the thesis fails, and the recovery was mostly a rebound.

Each quarterly release then updates the view. Transactions above 3% against tough comparisons would strengthen it. A big menu price increase, heavy discounting, or a negative quarter would weaken it. A rising two-year stack is the clearest sign that the base effect is fading and real demand is growing. Peer reports show whether a change is Starbucks' alone or industry-wide.

Common traps

• Revenue that hides units. Rising prices can keep revenue growing while customers leave, as at Apple in 2016–2018, PepsiCo's snacks in 2023, and Starbucks in 2024.
• Bought volume. Promotions can lift units briefly and then fade. Part of the bump is taken from rivals and part is borrowed from future weeks.
• Easy comparisons. Growth after a bad year is a rebound. Use two-year stacks.
• Calendar quirks. An extra week, a holiday that moves, or a leap year can add or remove a point or two of growth.
• New stores disguised as demand. Total volume can rise while each store serves fewer customers.
• Mix confusion. Cheap and premium units are not equal. A carload of coal and a carload of cars are both "one carload."
• Structural changes read as demand. Selling stores to franchisees or licensees can cut revenue while visits rise. Use per-store or system-wide measures.
• Treating falling volume as automatically fatal. Union Pacific grew profit for years while hauling less. Ask what is replacing volume as the engine, and how long it can last.
• Vanishing metrics. When a company stops reporting volume, as Apple did, ask why.

Checklist

When you study a company's volume growth, ask:

1. What is the right volume measure here, and is it reported consistently for at least three periods?
2. Has the definition, the store base, or the calendar changed in a way that breaks the comparison?
3. Is volume accelerating, steady, or slowing? What does the two-year stack say?
4. How much of revenue growth came from volume, and how much from price and mix? Does (1 + volume growth) × (1 + price growth) match the revenue growth?
5. Is volume rising at existing locations, or only through new stores and acquisitions?
6. Is growth broad across regions, products, and customer groups?
7. Was any volume bought with discounts that will end, or pulled forward from future periods?
8. How does the trend compare with peers and the industry cycle?
9. How close is the business to full capacity, and what will more volume cost?
10. If volume is falling, what is driving profit instead, and how long can that last?

Key takeaways

• Volume measures activity in units rather than dollars. Since revenue = volume × price, the same revenue growth can hide very different businesses.
• Falling volume hidden by rising prices is a common early warning sign. Rising volume at full price is strong evidence of real demand.
• Research shows that counts of activity, from parking-lot cars to filled airline seats, help predict sales and explain share prices, often before the financial results arrive.
• Judge volume on direction, breadth, and honesty (whether price is hiding or buying the trend).
• Separate new locations from existing ones, check the comparison base and the calendar, and compare with peers.
• Falling volume is a question, not a verdict. A company with pricing power can grow profit for a while on less volume, but the volume trend shows how much the story depends on price and cost cuts.

Data needed to support this thesis

To build and keep updating a volume growth thesis, you need the following data.

Required

Data	What it tells you	Where to find it
Operating metrics	The right measure of activity for this business, such as units shipped, customer visits, or active users	Results releases, earnings calls, and the "key metrics" section of annual reports (Form 10-K)
Volume history	Whether activity is speeding up or slowing, measured the same way over time	The same metric across at least three comparable periods of filings and releases

Helpful

Data	What it tells you	Where to find it
Industry data	Whether the whole market is growing or shrinking	Trade associations, government statistics, and peer results releases
Capacity data	How much room there is to grow before new stores, plants, or planes are needed	Store, seat, or plant counts in filings and releases
Customer metrics	Whether volume comes from more customers or more activity per customer	Results releases, earnings calls, and investor presentations

The comparisons must be like-for-like. If a company changes how it counts a unit, changes which stores count as "comparable", adds an extra week to its year, sells stores to licensees, or stops reporting a metric, adjust the history before comparing growth rates.

The thesis is usually updated by three kinds of events: earnings releases (new volume figures), customer metric updates (new counts of customers or activity), and industry data (market-wide activity). A typical horizon for checking the thesis is about a year and a half, long enough for easy comparisons to wash out and for a full year of new data to arrive.

In the Starbucks example, the required data was comparable transactions and average ticket for each period from fiscal 2018 onward, taken from the company's earnings releases.[^29][^30][^32][^31][^34][^35] The helpful data was store counts for capacity, and the McDonald's and Chipotle results as an industry check.[^18][^35][^37] The example used no formal industry statistics and no separate customer metrics. The Union Pacific case used carloads, revenue, and operating ratios from its annual reports, plus the regulator's findings on service.[^21][^22][^23][^24][^27]

---

[^1]: Apple Inc., Form 10-K for fiscal 2018: iPhone net sales $166,699 million versus $141,319 million; units 217,722 thousand versus 216,756 thousand. Revenue per iPhone is calculated. SEC filing. [^2]: Apple Inc., Form 10-K for fiscal 2016, showing iPhone net sales of $136,700 million, $155,041 million, and $101,991 million and units of 211,884 thousand, 231,218 thousand, and 169,219 thousand for fiscal 2016, 2015, and 2014, and attributing the 2016 decline primarily to "a lower rate of iPhone upgrades." SEC filing. [^3]: Apple, "The future is here: iPhone X" (September 12, 2017). Apple Newsroom. [^4]: AppleInsider, "Apple will no longer report iPhone, Mac and iPad unit sales" (November 1, 2018). AppleInsider. [^5]: The Motley Fool, "Why Apple Stock Plunged 18.4% in November" (December 11, 2018). Motley Fool. [^6]: Tim Cook, "Letter from Tim Cook to Apple investors" (January 2, 2019). Apple Newsroom. [^7]: Associated Press, "Stocks take a beating after Apple reports iPhone sales slip" (January 3, 2019). PBS NewsHour. [^8]: Apple Inc., Form 10-K for fiscal 2019: iPhone net sales $142,381 million, down 14%. SEC filing. [^9]: Apple Inc., Form 10-K for fiscal 2021: iPhone net sales $191,973 million, up 39%; Services net sales $68,425 million. SEC filing. [^10]: Zsolt Katona, Marcus Painter, Panos N. Patatoukas, and Jean Zeng, "On the Capital Market Consequences of Big Data: Evidence from Outer Space," Journal of Financial and Quantitative Analysis (2025). Cambridge Core; figures from Berkeley Haas. [^11]: Asher Curtis, Russell J. Lundholm, and Sarah E. McVay, "Forecasting Sales: A Model and Some Evidence from the Retail Industry," Contemporary Accounting Research 31(2) (2014). Wiley; working paper. [^12]: Eli Amir and Baruch Lev, "Value-relevance of nonfinancial information: The wireless communications industry," Journal of Accounting and Economics 22 (1996), pp. 3–30. ScienceDirect. Findings taken from the published abstract. [^13]: Bruce K. Behn and Richard A. Riley, "Using Nonfinancial Information to Predict Financial Performance: The Case of the U.S. Airline Industry," Journal of Accounting, Auditing & Finance 14(1) (1999), pp. 29–56. SAGE. Findings taken from the published abstract. [^14]: Brett Trueman, M. H. Franco Wong, and Xiao-Jun Zhang, "Back to Basics: Forecasting the Revenues of Internet Firms," Review of Accounting Studies 6 (2001), pp. 305–329. Springer; abstract at University of Toronto. [^15]: Michael V. Marn, Eric V. Roegner, and Craig C. Zawada, "The power of pricing," McKinsey Quarterly (2003), based on S&P 1500 companies. McKinsey. [^16]: Tammo H. A. Bijmolt, Harald J. van Heerde, and Rik G. M. Pieters, "New Empirical Generalizations on the Determinants of Price Elasticity," Journal of Marketing Research (2005). University of Groningen. [^17]: Harald J. van Heerde, Sachin Gupta, and Dick R. Wittink, "Is 75% of the Sales Promotion Bump Due to Brand Switching? No, Only 33% Is," Journal of Marketing Research 40(4) (2003), pp. 481–491. University of Groningen. [^18]: Chipotle Mexican Grill, results for the quarter ended June 30, 2026. SEC filing. [^19]: PepsiCo, Form 10-K for fiscal 2023: Frito-Lay North America net revenue up 7% with unit volume down 1%, and the prior year's 53rd week reducing revenue growth by 2 percentage points; PepsiCo Beverages North America net revenue up 5% with unit volume down 5%. SEC filing. [^20]: PepsiCo, Form 10-K for fiscal 2024: Frito-Lay North America net revenue down 1%, unit volume down 2.5%, operating profit down 7%; PepsiCo Beverages North America net revenue up 0.5%, unit volume down 3%. SEC filing. [^21]: Union Pacific, Form 10-K for 2014: revenue carloads 9,625 thousand, freight revenues $22,560 million, net income $5,180 million, operating ratio 63.5%, average employees 47,200. SEC filing. [^22]: Union Pacific, Form 10-K for 2019: revenue carloads 8,346 thousand, freight revenues $20,243 million, net income $5,919 million, operating ratio 60.6%, average employees 37,500, and earnings per share up 6% "despite volume declines of 6 percent." SEC filing. [^23]: Union Pacific, Form 10-K for 2022: revenue carloads 8,169 thousand, freight revenues $23,159 million, net income $7.0 billion, operating ratio 60.1% versus 57.2% in 2021, with the deterioration "driven by inflation, operational inefficiency, and higher fuel prices." SEC filing. [^24]: Union Pacific, Form 10-K for 2025: revenue carloads 8,447 thousand, freight revenues $23,220 million, net income $7,138 million, operating ratio 59.8%. SEC filing. [^25]: Union Pacific, Form 10-K for 2016, reporting that coal shipments from the Powder River Basin declined 15% in 2015, citing low natural gas prices and high inventory levels. SEC filing. [^26]: Union Pacific, Form 10-K for 2018, describing the launch of Unified Plan 2020 and the start of precision scheduled railroading on October 1, 2018. SEC filing. [^27]: Surface Transportation Board, "STB Issues Notice of Public Hearing Regarding Union Pacific Embargoes" (November 22, 2022). STB. [^28]: Surface Transportation Board, "STB Issues Findings on Union Pacific Embargoes" (April 2024). STB. [^29]: Starbucks earnings releases for fiscal 2018, 2019, 2020, 2021, and 2022. [^30]: Starbucks earnings releases for fiscal 2023, 2024, and 2025, and for fiscal 2026 Q1, Q2, and Q3. [^31]: Starbucks, "Starbucks Reports Q4 and Full Fiscal Year 2025 Results" (October 29, 2025). Starbucks IR. [^32]: Starbucks, "Starbucks Reports Q4 and Full Fiscal Year 2024 Results" (October 30, 2024). Starbucks IR. [^33]: CBS News, "Starbucks stock soars after it ousts CEO Narasimhan, replacing him with Chipotle's Brian Niccol" (August 13, 2024). CBS News. [^34]: Starbucks, "Starbucks Reports Q1 Fiscal Year 2026 Results" (January 28, 2026). Starbucks IR. [^35]: Starbucks, "Starbucks Reports Q3 Fiscal Year 2026 Results" (July 29, 2026), quarter ended June 28, 2026. SEC filing. [^36]: Starbucks, "Starbucks Reports Q2 Fiscal Year 2026 Results" (April 28, 2026), quarter ended March 29, 2026. SEC filing. [^37]: McDonald's, results for the quarter ended June 30, 2026. SEC filing. [^38]: Starbucks, "Starbucks Is Back, Turning Momentum Into Long-Term, Sustainable Growth" (Investor Day, January 2026). Starbucks IR.