How the learning loop works

Overview

The Content Score Engine is how Postlore learns what works for your accounts: which times, days, hashtags, formats and caption lengths get you the most engagement. It keeps a separate picture for each workspace and each platform, and every AI suggestion uses it.

It starts on day one, from your existing posts

You don't have to publish through Postlore for weeks before it knows anything. When you connect a platform, Postlore imports your posts from the last 90 days (up to 100–200, depending on the platform) together with their engagement, and builds your patterns straight away.

• YouTube, TikTok, Pinterest, Reddit and Bluesky: imported on every plan, including Free.
• Facebook, Instagram, Threads, Google Business Profile, and LinkedIn: imported on paid plans.
• X: not imported yet. Postlore learns from the posts you publish there through Postlore.

If you already post regularly, you can have useful patterns within minutes of connecting.

It keeps learning from every post

After you publish, Postlore checks each post's performance at 24 hours, 48 hours and 7 days, and folds the results in. Recent posts count for more than old ones, so the picture follows your audience as it changes.

What it tracks

• Best posting hours and days, per platform, on your workspace's own clock (its timezone in Settings).
• Top hashtags, and which platform each works best on.
• Caption length and media type: short, medium or long; text, image or video.
• Post format and goal, such as questions, lists or announcements, and posts aimed at reach, conversation or sales.
• Language: for example, whether bilingual posts outperform single-language ones.
• Visual templates, music and brand colours used in Postlore-made visuals.

How confident is it?

Confidence grows with the number of posts that have engagement data, imported or published: low from 3, medium from 10, and high from 20. Until your own data is there, Postlore leans on trend signals from your market and industry.

How it improves your content

Every AI tool uses what Postlore has learned, on every plan. The idea generator leans towards the formats and topics that worked for you, Adapt writes with your strongest caption length in mind, and the hashtag tool prioritises your proven tags.

The detailed view (performance patterns in Insights, and the Readiness Score and best-time suggestions in Compose) is on the Creator plan and above. Free workspaces get a one-time 30-day full-analytics trial once they have data, and a one-time look at their score as it reaches each confidence level.

Voice fingerprinting and the learning loop

The Content Score Engine learns when and in what form to post. Voice fingerprinting goes a level deeper: it analyses the writing style of your top-performing posts (sentence length, emoji use, calls to action, questions) and applies that style to future AI output. The score engine works automatically; voice fingerprinting is something you switch on in Brand settings.