Time series and cross sectional analysis: when to use each

Time series and cross sectional analysis answer different questions from the same business file: what changed over time, and how units differ in one period. A branch can lead the current sales ranking while showing no growth. Read both views before deciding where to investigate or allocate operating attention.

Choose the question before the comparison

A time series keeps the business unit and measure consistent while the period changes. A cross-section keeps the period and measure consistent while the unit changes. The distinction concerns the observations you compare, rather than the chart you choose.

Time ordering matters. The NIST introduction to time series analysis explains that observations over time can contain trend, seasonality, and autocorrelation. A short monthly table establishes a comparison; it does not establish a stable long-term pattern.

For the broader treatment of changes through time, see time series analysis. The task here is choosing the comparison structure before interpreting it.

Start with one consistent branch export

Consider an illustrative US industrial distributor reviewing January through March 2026. Its report contains net sales after credits and productive labor hours for each branch-month. All amounts are USD, all branches were open throughout, and the measure uses the same posting-date policy in each month.

Store actual month-start dates in the workbook, even though the table displays month names. A label such as March without a year will eventually combine unrelated periods. Use branch IDs rather than names if branch names can change.

With the header in row 1, columns A through D hold branch, month, net sales, and hours. Keep one row per branch-month. If your export contains invoice lines, aggregate to this level and retain a reconciliation to the invoice total before making either comparison.

Work the time series first

Filter the export to Atlanta and preserve the month order. January sales are the starting point and March sales the endpoint.

Endpoint growth = (Ending-period sales - Starting-period sales) / Starting-period sales

In Excel, the calculation on these exact rows is:

=(C4-C2)/C2

Atlanta grew 20.0% between the endpoints. That does not mean every branch grew at that rate, and it is not a monthly compound growth rate.

Summing all branches within each month gives the company time series:

Company endpoint growth is (375000-330000)/330000, or 13.6%, rounded to one decimal place. The monthly totals sum to the quarter's $1,050,000. That is the first reconciliation: changing the grouping must preserve the source total.

Work the cross-section from the same rows

Now hold March fixed and compare branches. Sales share uses company March sales as its denominator. Sales per productive hour uses each branch's own March hours.

Denver leads in sales dollars; Atlanta leads in sales per productive hour. Neither ranking establishes profitability, service quality, or management effectiveness. Branches can differ in product mix, delivery territory, and customer requirements.

If F2 holds a branch ID and G1 a March 1, 2026 date, this Excel formula returns that branch's March sales from the source layout:

=SUMIFS($C$2:$C$10,$A$2:$A$10,F2,$B$2:$B$10,$G$1)

Microsoft documents SUMIFS as summing values that meet multiple criteria. Here the criteria are branch and month. Covirage is not affiliated with Microsoft.

The total rate is 375000/3500, or $107.14 per hour, rounded to cents. Divide summed sales by summed hours; averaging branch rates gives each branch equal weight regardless of its hours.

Read position and direction together

The same source rows give Denver endpoint growth of (150000-150000)/150000 = 0.0%. Phoenix grows (105000-80000)/80000 = 31.25%. Denver is the largest March branch while Phoenix has the fastest endpoint growth. Those findings support different conversations.

Operations owns the explanation of branch capacity and service constraints. Finance owns the consistent sales definition and reconciliation. The branch manager supplies the local facts needed to interpret the gap. The report identifies where to ask; it does not supply a causal explanation by itself.

External comparisons also need a defined reference period and population. For example, the Census Bureau's County Business Patterns publishes annual subnational data about establishments, employment, and payroll. That can provide operating context, but its measures are not substitutes for the distributor's branch sales.

For statement-level peer comparisons, use cross-sectional financial statement analysis. For an operational measure grouped across both suppliers and periods, see the on-time-in-full trend by supplier and site.

Where it goes wrong

• Changing branch coverage. An acquisition or opening moves company sales even when existing branches are unchanged. Show a consistent-branch view beside the reported total.
• Mixing periods. A partial month should not be ranked against completed months. Apply the same close cutoff to every branch.
• Using size as performance. Sales dollars establish scale. Add a relevant denominator, then investigate differences in operating conditions.
• Treating a short run as a forecast. These observations show changes already recorded. More history and separate validation are needed before making a forecast.
• Losing the total. Both summaries must reconcile to their selected source rows. A missing branch or duplicate month undermines the comparison before interpretation begins.

Compare your own branch file in Covirage

Bring a branch-by-month sales file with consistent period labels and any operating denominators you want to compare. Covirage maps the columns, and its tools compute and check the totals, changes, and rates; the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See self-service analytics for the supporting workflow.