What’s the question?
Revenue seasonality is the share of a fiscal year’s revenue that lands in each of that year’s four quarters. A business with no seasonality books 25 percent in each. A retailer that clears most of its inventory over the holidays does not, and neither does a tax-preparation software company whose customers all show up in April.
The consequence appears whenever two consecutive quarters are compared. A sequential revenue decline can mean the business shrank, or it can mean the calendar turned, and the number itself does not say which. Any screen ranking companies on quarter-over-quarter revenue growth treats both cases identically.
A second complication is easy to miss. Fiscal quarters are not calendar quarters: Apple’s fiscal Q1 ends in December while Costco’s fiscal Q4 ends in August, so grouping companies by the label "Q4" mixes the holiday shopping season with late summer.
The approach
The sample starts with 29 S&P 500 companies spanning retail, staples, software, hardware, industrials, healthcare and energy, over July 2015 to December 2025.
- Pull quarterly revenue, collapsing any records that describe the same reporting period into a single row.
- Derive the fiscal quarter position from the data instead of the calendar. A quarter whose period end matches an annual period end is that company’s fiscal Q4, and the three before it are Q3, Q2 and Q1. Calendar month is never used, which matters because only 12 of these companies close their books in December.
- Keep companies whose quarters run consecutively, allowing gaps of 75 to 130 days for 52/53-week calendars and for Costco’s 16-week fourth quarter.
- Require each fiscal year’s four quarters to reconcile to the company’s own reported annual revenue within 1 percent. Restatements after a divestiture leave some years unreconciled; those years drop, and a company needs eight reconciling years to stay in.
- Take each quarter’s share of its fiscal year, average by quarter position, and use the highest minus the lowest average as the seasonality statistic.
That leaves 28 companies and 269 fiscal years; Johnson & Johnson reconciles in seven and drops. The derived positions agree with the fiscal period recorded on each filing for 1,176 of 1,177 quarters, and 276 of the 283 complete fiscal years reconcile to within 0.01 percent. Walmart’s year ending 31 January 2025 sums to 674,538 against a reported 674,538, and Apple’s year ending 27 September 2025 sums to 416,161 against 416,161.
Code
import pandas as pd
import xfinlink as xfl
xfl.set_api_key("YOUR_API_KEY") # free at https://xfinlink.com/signup
TICKERS = ["HD", "LOW", "TGT", "BBY", "ROST", "NKE", "WMT", "COST", "KO", "PG",
"CL", "KMB", "MSFT", "ADBE", "ORCL", "CRM", "INTU", "AAPL", "HON",
"CAT", "UNP", "EMR", "JNJ", "ABT", "MRK", "UNH", "XOM", "CVX", "COP"]
START, END = "2015-07-01", "2025-12-31"
qtr = xfl.fundamentals(TICKERS, period_type="quarterly", start=START, end=END, fields=["revenue"])
ann = xfl.fundamentals(TICKERS, period_type="annual", start=START, end=END, fields=["revenue"])
# Fiscal quarter POSITION, derived from the data: a quarter whose period end matches
# an annual period end is that company's fiscal Q4. Calendar month is never used.
frames = {}
for t, g in qtr.groupby("ticker"):
g = g.sort_values("period_end").reset_index(drop=True)
gaps = g["period_end"].diff().dt.days.dropna()
ends = ann.loc[ann["ticker"] == t, "period_end"]
q4 = [i for i, d in enumerate(g["period_end"]) if (ends - d).abs().dt.days.min() <= 7]
if len(g) < 38 or not gaps.between(75, 130).all() or g["revenue"].le(0).any():
continue
if len(q4) < 5 or len({i % 4 for i in q4}) != 1:
continue
g["fq"] = (g.index - q4[0] + 3) % 4 + 1
frames[t] = g
q = pd.concat(frames.values(), ignore_index=True)
# Complete fiscal years that reconcile to the company's reported annual revenue.
q["block"] = q.groupby("ticker")["fq"].transform(lambda s: (s == 1).cumsum())
q["n_in_block"] = q.groupby(["ticker", "block"])["fq"].transform("size")
q["fy_rev"] = q.groupby(["ticker", "block"])["revenue"].transform("sum")
q["fy_end"] = q.groupby(["ticker", "block"])["period_end"].transform("max")
fy = q[q["n_in_block"] == 4].merge(
ann[["ticker", "period_end", "revenue"]].rename(
columns={"period_end": "fy_end", "revenue": "annual"}),
on=["ticker", "fy_end"], how="inner")
fy["recon"] = (fy["fy_rev"] - fy["annual"]).abs() / fy["annual"]
fy = fy[fy["recon"] <= 0.01]
fy = fy[fy.groupby("ticker")["block"].transform("nunique") >= 8]
fy["share"] = fy["revenue"] / fy["fy_rev"]
prof = fy.pivot_table(index="ticker", columns="fq", values="share", aggfunc="mean")
prof.columns = [f"Q{c}" for c in prof.columns]
prof["spread"] = prof.max(axis=1) - prof.min(axis=1)
prof["peak"] = prof[["Q1", "Q2", "Q3", "Q4"]].idxmax(axis=1)
prof = prof.sort_values("spread", ascending=False)
# The quarter right after the peak, read sequentially and year-over-year.
q = q[q["ticker"].isin(prof.index)].copy()
q["qoq"] = q.groupby("ticker")["revenue"].pct_change(1)
q["yoy"] = q.groupby("ticker")["revenue"].pct_change(4)
for t in prof.index:
nxt = int(prof.loc[t, "peak"][1]) % 4 + 1
s = q[(q["ticker"] == t) & (q["fq"] == nxt)].dropna(subset=["qoq", "yoy"])
print(f"{t:6}{prof.loc[t, 'spread']:8.1%}{s['qoq'].median():9.1%}{s['yoy'].median():9.1%}"
f"{int(((s['qoq'] < 0) & (s['yoy'] > 0)).sum()):>6}/{len(s)}")Full script with formatting and visualisation: quarterly-revenue-seasonality-python.py
Output
Companies screened: 29 in final sample: 28 fiscal years used: 269
Share of fiscal-year revenue by fiscal quarter position
Q1 Q2 Q3 Q4 spread peak yrs sector
INTU 15.9% 20.6% 44.6% 19.0% 28.7% Q3 10 Information Technology
BBY 21.2% 21.9% 22.9% 34.1% 12.9% Q4 9 Consumer Discretionary
AAPL 32.1% 23.3% 21.1% 23.4% 11.0% Q1 10 Information Technology
COST 22.4% 23.1% 22.7% 31.7% 9.3% Q4 10 Consumer Staples
TGT 22.6% 23.6% 23.8% 30.0% 7.3% Q4 9 Consumer Staples
LOW 24.4% 28.8% 24.2% 22.5% 6.3% Q2 9 Consumer Discretionary
ROST 22.4% 24.4% 24.7% 28.5% 6.1% Q4 9 Consumer Discretionary
ORCL 23.1% 24.3% 24.7% 27.8% 4.7% Q4 10 Information Technology
EMR 22.8% 24.6% 25.3% 27.3% 4.6% Q4 8 Industrials
CRM 23.0% 24.3% 25.5% 27.2% 4.3% Q4 9 Information Technology
HD 23.7% 27.9% 24.7% 23.7% 4.2% Q2 9 Consumer Discretionary
MSFT 22.8% 25.7% 24.4% 27.0% 4.1% Q4 10 Information Technology
MRK 24.9% 24.9% 26.7% 23.5% 3.2% Q3 10 Health Care
WMT 23.7% 24.8% 24.6% 26.8% 3.1% Q4 9 Consumer Staples
CAT 23.5% 25.3% 24.7% 26.5% 2.9% Q4 10 Industrials
COP 26.0% 23.2% 25.4% 25.4% 2.9% Q1 10 Energy
ADBE 23.8% 24.5% 25.2% 26.6% 2.7% Q4 10 Information Technology
UNP 24.1% 24.1% 25.0% 26.7% 2.6% Q4 10 Industrials
ABT 24.0% 24.6% 25.2% 26.3% 2.3% Q4 10 Health Care
CVX 23.9% 24.3% 26.1% 25.8% 2.2% Q3 10 Energy
KMB 25.6% 25.4% 25.6% 23.4% 2.2% Q1 10 Consumer Staples
KO 24.0% 26.1% 25.7% 24.1% 2.1% Q2 10 Consumer Staples
PG 25.3% 25.8% 24.1% 24.8% 1.7% Q2 10 Consumer Staples
XOM 24.1% 24.4% 25.7% 25.7% 1.6% Q3 8 Energy
NKE 25.6% 25.1% 24.5% 24.8% 1.1% Q1 10 Consumer Discretionary
UNH 24.6% 24.8% 25.1% 25.5% 0.9% Q4 10 Health Care
HON 24.5% 25.3% 25.3% 25.0% 0.8% Q3 10 Industrials
CL 24.8% 24.8% 25.2% 25.3% 0.5% Q4 10 Consumer Staples
Quarter after the peak: sequential read vs year-over-year read
qtr QoQ YoY gap(pp) false drops
INTU Q4 -56.0% 14.6% 70.7 9/10
BBY Q1 -37.1% -0.9% 36.2 4/9
AAPL Q2 -24.1% 4.6% 28.7 6/9
COST Q1 -22.0% 8.2% 30.2 10/10
TGT Q1 -22.9% 3.4% 26.3 7/9
LOW Q3 -15.6% 3.0% 18.5 7/10
ROST Q1 -13.7% 5.8% 19.5 6/9
ORCL Q1 -11.8% 4.9% 16.7 10/10
EMR Q1 -11.9% 2.8% 14.7 7/10
CRM Q1 1.2% 24.3% 23.1 3/9
HD Q3 -11.5% 5.8% 17.3 9/10
MSFT Q1 -2.1% 13.2% 15.3 7/10
MRK Q4 -5.6% 5.2% 10.8 5/10
WMT Q1 -7.8% 2.6% 10.4 9/9
CAT Q1 -4.4% 4.7% 9.1 4/9
COP Q2 -7.9% 14.9% 22.9 2/9
ADBE Q1 3.6% 18.8% 15.2 0/9
UNP Q1 -2.0% 2.5% 4.5 3/9
ABT Q1 -2.7% 4.0% 6.7 6/9
CVX Q4 -0.9% 3.0% 3.9 2/10
KMB Q2 -0.8% 0.4% 1.2 5/9
KO Q3 -3.2% 2.1% 5.3 5/10
PG Q3 -6.4% 3.5% 9.9 7/9
XOM Q4 0.0% 1.0% 1.0 2/10
NKE Q2 0.2% 5.5% 5.3 5/10
UNH Q1 6.0% 9.4% 3.4 0/9
HON Q4 3.0% -1.2% -4.2 0/10
CL Q1 1.9% 5.5% 3.6 1/9What this tells us
The range across 28 large, mature, profitable companies runs from 0.5 percentage points to 28.7. Colgate-Palmolive divides its year almost evenly into quarters of 24.8, 24.8, 25.2 and 25.3 percent, while Intuit puts 44.6 percent of its revenue into fiscal Q3, the quarter ending in April, because United States personal tax returns fall due in the middle of that month. Those two companies need different reading rules, and nothing on the face of a revenue series announces which rule applies.
The fiscal-versus-calendar problem is not academic. Fifteen of these companies show their largest quarter in fiscal Q4, and those fourth quarters end in eight different calendar months, from Costco in August through Oracle in May to Target in February. Sorting by the label "Q4" groups Costco’s summer with Target’s holiday season, while sorting by calendar month splits Apple’s December quarter away from Best Buy’s January-ending one even though both capture the same shopping weeks. The flat end of the table follows the business models: Colgate, Honeywell, UnitedHealth and Nike all sit under 1.2 points of spread, because consumables, service contracts, insurance premiums and wholesale shipments accrue at close to constant rates.
The final table prices the error. For the 11 most seasonal names, 78 of 105 post-peak quarters showed a sequential revenue decline while the same quarter a year earlier was lower, meaning the decline reversed sign under a year-over-year read; Costco and Oracle did this in every observation. The median gap between the two growth readings across those companies is 23.1 percentage points, against 3.9 points and a false-decline rate of 36 of 104 for the 11 flattest names. Apple shows the mechanism: its fiscal Q2 carries a median sequential change of negative 24.1 percent against a median year-over-year change of positive 4.6 percent, and in the March 2025 quarter revenue fell 23.3 percent from the December quarter while rising 5.1 percent against March 2024. Both numbers are correct; only one describes the business. Year-ago quarters also serve as the naive benchmark in a comparison of how forecastable quarterly revenue and earnings are.
So what?
Compute the seasonal profile before comparing any two quarters and store the spread alongside the ticker. Ten years of quarterly revenue produce a single number per company that settles which comparison is legitimate.
Above roughly 4 points of spread, sequential revenue growth carries no usable information about the business; the 12 companies past that threshold here produce a phantom decline in most post-peak quarters, so use year-over-year growth instead. Below about 2 points, sequential growth is sound and reacts faster, which matters for flat names where a genuine turn otherwise stays buried for three quarters.
Any screen ranking a universe on quarter-over-quarter revenue growth is close to a ranking of who is furthest past their seasonal peak. One adjustment fixes it: divide each company’s quarter by the average share that quarter position carries in its own fiscal year, which puts Costco’s August quarter and Colgate’s March quarter on the same footing. Derive that position from period-end dates, never from calendar month, or the correction lands on the wrong quarter for half the names.
Built with xfinlink — free financial data API for Python.