Efficient use tactics

Proven tips for spending credits smartly, training loom fast, and turning score into real results.

last updated: August 19, 2026

While generating

+1. Variant strategy: Short/Medium cost 1 credit — generate 3–4 variants in different tones on the same topic, pick the best. Total cost is usually 3–4 credits.
+2. Read the score breakdown: on a low score it writes which item is missing (no hook, no question...) — change the tone/type before regenerating.
+3. Give a source: instead of generating from a bare topic, give a tweet or article link — the output becomes concrete, and the specificity points come on their own.

+4. The first week is feedback week: approve/reject/edit — every signal is written to memory. State the reason with one click when rejecting; learning speeds up noticeably.
+4b. Leave free-form notes: use "✎ feedback" on an approval card to coach loom in your own words ("too flat", "don't end sentences with a period") — teach only, or have it fix the draft with the same note. Check "make it a permanent rule" to save the note as a HARD RULE applied to every loom generation (managed in settings → persona).
+5. Build a source mix: 2–3 high-signal X accounts + 1–2 news sources. A single source produces single-type drafts.
+6. Walk your approval queue for 2 minutes every day — steady feedback (approve/reject/edit) is the fastest way to sharpen loom's accuracy.

+7. Schedule to your best time: loom learns your best hours from past performance; scheduled posting consistently beats instant posting.
+8. Wait for the measurement: every post is measured from X 24 hours and 7 days later — the calibration on the performance page shows which content type works for YOU; generation is shaped by this data too.
+9. Keep the pace human-scale: X can quietly restrict reach when it sees high volume — especially on smaller accounts (40+ posts a day, repeated replies to the same account, replies with links). tweetloom's limits are deliberately conservative for this reason — few but visible posts always beat many but buried ones.

tipNone of these tactics are theory — they all rest on measured mechanisms in the product.

You've read it — now try:put it into practice