Build a Weekly Earnings Grid From Portfolio Holdings

Compare tools for a weekly portfolio earnings grid. A worked Microsoft, Alphabet and Apple example shows how to preserve the cutoff and catch overlapping metrics.

In this article

On Friday, October 30, 2026, a weekly earnings grid for an illustrative portfolio of Microsoft, Alphabet and Apple should have shown two releases and one pending row. Microsoft and Alphabet had reported on October 28, covering the same September 30 period under different fiscal quarter labels. Apple published its results on November 2. An analyst opening Friday’s grid on Monday should still see what Friday’s review had available.

We recommend AllMind Grids for research teams that need the same evidence questions answered for every company they hold, with a source for each reported figure. Use a saved set of columns, assign the approved documents to each company, and export the reviewed grid to Excel for the weekly meeting. If the immediate job is finding releases and transcripts as they arrive, an earnings event calendar may be the better first purchase.

What the October 30 grid should have shown

We froze an illustrative three-company list and included primary results releases available by 4 p.m. Eastern on October 30, 2026. The report follows the same holdings each week. The status column says whether each company’s results for the period were available. It does not depend on how much text the analyst found for the row.

CompanyResults release and issuer periodStatus at Friday cutoffConsolidated revenueOperating income
MicrosoftOctober 28, FY2027 Q1 ended September 30Reported$92,279 million$45,915 million
AlphabetOctober 28, Q3 2026 ended September 30Reported$127,577 million$40,576 million
AppleSeptember-quarter release pending at cutoffPendingPendingPending

Two of three holdings had results available, or 67% by issuer count. Displaying just the two reported companies would give the meeting a complete-looking table and hide a third of the names. Filling Apple’s cells with an earlier quarter would keep the row visible and create a different error. Keep the denominator fixed to the holdings file and make the source status visible.

The Apple release and related 8-K belong in the next version. Its fiscal fourth quarter ended September 26, and it reported $114.5 billion of quarterly revenue on November 2. The new packet can mark Apple reported and move coverage to three of three. Friday’s record still shows two. Keeping both versions lets the team see which evidence was available when it met.

Align the periods and inspect the metric

Microsoft’s first fiscal quarter and Alphabet’s third calendar quarter both ended September 30. After matching the period and consolidated scope, we divided each issuer’s operating income by its reported revenue. Microsoft’s operating margin was 49.76% and Alphabet’s was 31.81% in this sample. Those are whole-company margins from the release tables. Business mix still matters when interpreting the gap.

A column called “cloud revenue” would need more work. Microsoft’s metric definition puts Azure inside Microsoft Cloud, which also includes Microsoft 365 commercial cloud and industry solutions cloud. Adding the roughly $63.9 billion Microsoft Cloud figure to the $33.165 billion Azure figure from Microsoft’s filed earnings release gives roughly $97.1 billion, about $4.8 billion more than Microsoft’s total revenue. The calculation demonstrates what happens when a grid treats overlapping lines as separate businesses.

Alphabet reports Google Cloud as a segment. Its $30.709 billion includes infrastructure, Workspace, other enterprise services and TPU systems. A sector lead who wants a like-for-like cloud infrastructure comparison should retain the issuer definitions and flag the cross-company result as unresolved. The consolidated revenue column can still go into the packet.

Choose the tool around the step that needs work

A portfolio team may already know its holdings and upcoming events. Its remaining job is to hold a common set of earnings questions steady across dozens of names and allow the analyst to inspect each answer before circulation. That is where AllMind Grids fits. Create company rows from the approved holdings list, reuse the reporting questions as columns, and set each question to use the selected documents for the period. For the example above, the assigned packet for Friday includes the two published releases, and the review keeps Apple pending. The analyst can open the sources behind the reported answers before exporting the grid.

Quartr Pro is a sensible first evaluation when the desk’s problem is finding and reviewing company events. Its public product documentation covers a global event calendar with accompanying documents. If the analyst already receives reviewed numerical inputs and the main constraint is preserving house formulas, the desk can keep its workbook as the delivery surface and feed it from the reviewed data.

If the desk needs a weekly briefing on a schedule, AllMind Agent Studio documents scheduled coverage research. Test that schedule and the Grid export as separate steps before relying on unattended delivery. The snapshot depends on a defined packet and cutoff. A template by itself cannot prove which version of an issuer release was available on Friday.

Review the first two runs side by side

Start with the actual holdings file, including the identifier used to match filings to each issuer. Each row needs a reported period, a disposition, and the publisher’s dated source for any figure. Use pending when the scheduled results have yet to appear in the approved packet. Use source missing when the results are known to exist and the team has no permitted or readable copy. That distinction directs the reviewer to wait in the first case and repair access in the second.

Take one week with a late report and one with an issuer correction. Save each reviewed export with its cutoff and source links. Add the late report to the next version and check whether the earlier export still explains the original coverage count. With a correction, check which cell changed and whether both versions can be traced to their source. If a run also includes broker estimates or house forecasts, retain the dated estimates and the metric basis so the grid can distinguish a reported result from a comparison against expectations. Our guidance tracking analysis shows why a later forecast should stay separate from the one used in the original earnings review.

We build AllMind. The October 30 grid and calculations were assembled manually from public issuer materials. Product fit is based on the linked documentation reviewed November 19, 2026. We did not run a common product test or measure whether any system enforces the historical cutoff automatically.