Telegram Ads change journal: before and after

After editing ad text, it is tempting to credit every CPA movement to that edit. In Adtally, the Journal keeps changes and hypotheses alongside data so each review has a date, a scope and a stated question.

Record the hypothesis first

Create an entry, select its type and choose an ad or campaign. Enter the date and time in UTC, a description and a review date. Hypotheses also have a status and conclusion.

State a testable expectation. A fictional example: “After changing the first sentence, I expect CTR to increase at comparable impression volume; I will check CPA separately.” Make the edit in Telegram Ads and record what you changed. The dashboard only reads advertising data.

Distinguish a record from a detected change

When a fresh ad list loads, the journal compares observed settings with their previous values. It can flag changes to text, budget, status, targeting and other tracked fields.

That is the detection time, not the exact edit time. After several days away, a change is only known to have occurred between two observations. The first load establishes a baseline rather than reconstructing earlier edits. Use a manual record when you know the exact time.

Wait for two complete windows

Automatic comparison requires seven complete days before an entry and seven after. An edit made during a day excludes that day; boundaries use UTC.

For example, an entry at 14:00 on October 10 compares October 3–9 with October 11–17. The second window finishes at 00:00 UTC on October 18. Load a date range containing both windows: incomplete statistics leave the comparison unready.

Gaps and unavailable ads prevent evaluation. Different goals and conversion events are not combined into CPA or CVR. See the missing data guide.

Separate observation from cause

A lower CPA after an edit can coincide with changes to budget, audience or the selected ads. Before-and-after comparison shows a relationship in time; it does not establish the effect of the edit.

Change one parameter at a time, record accompanying changes and read spending, actions and CTR together. Use equal periods and recommendations as additional checks. “Confirmed” is your recorded conclusion rather than automatic proof.

Entries appear on charts and can be included in a client report XLSX. The journal is local, with up to 500 entries and 2,000 observed ads. Keep important conclusions separately.