Skip to content

Slots that have done better

How the next-best-slot card decides one day and part of the day did measurably better, and the three floors it has to clear.

Updated

The "Slots that have done better" card on your calendar names one day of the week and one part of the day where your own published posts have recorded a higher median engagement than your overall median. It is a record of what already happened, never a forecast.

It refuses far more often than it answers, on purpose. Three separate floors have to be cleared before it will name anything, and when it cannot it tells you exactly which floor was missed.

BEFORE YOU START

  • Posts published through Social Studio, with a figure held against the post's viral signal. That is a narrower thing than it sounds and it is the requirement people miss: it is not the same as a row in your reporting. Publishing records the signal with no figure on it, so a post you published and later reported figures for still counts as unmeasured here. The last bullet under "What this will not do" says which path does reach it.
  • A recorded publication time on those posts. A post whose scheduled time was never recorded is excluded and counted separately, because the moment a worker happened to finish is not when the post went out.
  • Nothing else. Reading this costs no credits and needs no permission beyond membership of the organisation.

STEPS

  1. Open the calendar for your organisation.
  2. Find the card headed "Slots that have done better", under the eyebrow "Evidence". It sits beside a different card about your posting cadence, which answers availability rather than performance.
  3. If it names a slot, read the two figures under it: how many measured posts sit in that slot, and the median engagement there. It is a median, never an average, so one unusually far-travelling post cannot decide the recommendation.
  4. Read "Next dates that slot comes round". You get dates and a part of the day, never a time. Picking the hour inside that band is yours.
  5. If it refuses, read the three requirements it lists. Each says exactly what is missing.

WHAT YOU SHOULD SEE

Either a named day and part of the day, with the evidence behind it and the next three dates it comes round, or the heading "Not enough evidence yet" with the reason and the three floors spelled out.

WHAT THIS WILL NOT DO

  • It does not suggest a free slot. That is a different card and a different question. This one only reports which slots your posts have measurably done better in.
  • It does not name an hour. The figures behind it are recorded daily, so it cannot tell 09:05 from 09:55, and the hour it puts on a date is a position inside the band rather than a recommendation.
  • It does not name a slot that merely tops a flat set of results. If no slot beats your overall median, it says so and tells you to post when it suits you.
  • It is not on the mobile app. The phone shows the posting-cadence card only.
  • It does not read your reporting figures, and nothing in the product carries them across for you. Figures sent to the workspace report-metrics API land in reporting and stop there. The code that would copy them onto the matching viral signal exists and is never run, so however many figures you report, this card's sample stays where it was. The one path that does raise it today is recording the figures as a viral signal yourself, by posting them to /api/v1/viral/signals with the same platform and external reference as the published post; that updates the signal publishing already created, and the post then counts as measured.

IF IT DOES NOT WORK

  • "We can see N of your posts with a recorded figure and a publication time, and we need 12" means the sample is too small. Publishing more adds posts, but it does not add MEASURED ones: publishing records each post's viral signal with no figure on it. Reporting the figures does not raise this count either, for the reason given in the last bullet above. Recording them as viral signals does.
  • "Your posts are spread thinly across the week" means no single day and part of the day has the three measured posts it needs.
  • "No slot has done measurably better than your others." means the results are flat. Naming one anyway would be reading a pattern into noise.
  • A count of posts "published with a figure but no recorded publication time" explains why the sample is smaller than your published-post count. Those posts say nothing about timing.
  • "The request did not complete, so nothing is known about which slots have done better." means the card failed to load, which is not the same as it having nothing to say.

COMMON QUESTIONS

Why a median and not an average?

Because one post that travelled unusually far would otherwise decide the whole recommendation.

Why does a slot at the top of the spread list not get named?

Because three measured posts in a slot is the floor, and one or two do not make a pattern. The spread list is shown so you can see the shape of your data, not as a shortlist.

  • Planning your week

    Building a week from your own calendar and posting history, and why no suggestion is ever a predicted best time.

  • Planning posts you publish yourself

    Manual slots hold a time for a network Social Studio cannot publish to, and record where you posted it.