How a mid-size asset manager halved the time from earnings release to client note

Kestrel Capital is a fictional asset manager used on this site for illustration. The figures below are invented to show how a case study is structured and do not describe any real firm.

The problem

Kestrel's equity research team of nine analysts covered around 120 companies. During reporting season, each analyst faced a queue of results releases, often several in one morning. A first note to portfolio managers and clients typically took four to five hours, by which point the share price had already reacted and the commentary was describing a move rather than anticipating one.

Analysts said the delay was not in forming a view but in assembling the numbers. Each release meant manually transcribing figures from a press release into a spreadsheet, recomputing the model, checking against consensus and formatting a note. Errors were common under time pressure, and a handful of notes each season had to be reissued.

What they did

The head of research began by timing the process. Roughly 60 per cent of the elapsed time was spent on data entry and formatting; the analytical work took about an hour.

The team built a small data pipeline that pulled headline figures from the results release as soon as it was published, matched them to the analyst's model and produced a one-page comparison against the analyst's forecast and consensus. The pipeline flagged any line where the reported figure differed from the model by more than a set threshold, which caught both surprises and transcription errors.

The note itself was templated. A first note now had a fixed structure: headline versus expectations, guidance change, three bullet points on what mattered, and a placeholder for the analyst's view. The analyst filled in the view and the bullets; the numbers were already in place.

A second, fuller note followed after the analyst call, on the old free-form format. The split let the desk publish something useful within the first hour without committing to a full analysis before the call.

The compliance team reviewed the template so that the standard disclosures and disclaimers were built in, removing a checking step that had previously happened at the end.

Results

Across the two reporting seasons after launch, the median time from release to first note fell from about four and a half hours to just under two. The share of covered companies with a note published on the day of the release rose from 58 per cent to 91 per cent.

Data errors caught after publication fell by 72 per cent, which the head of research attributed to the pipeline's threshold check. Portfolio managers reported that the shorter first notes were more useful than the longer ones had been, because they arrived while a decision was still open.

What others can take from it

Time the process before changing it. The team's assumption was that analysis was the bottleneck; the measurement showed it was clerical work.

Separate the fast product from the thorough one. Trying to make a single note both quick and complete had produced one that was neither.

Build the checks into the tool. The threshold flag caught errors that a tired analyst at hour four had not, and it did so before publication rather than after.

Involve compliance early. Getting disclosures into the template removed a step and a source of friction between the desks.