Building an "Instagram + TikTok Hashtag Search → Auto-DM Generation" Pipeline with n8n × Apify × Claude

First up, you're ballparking Apify costs, but where's the actual receipts from running this in production? Then, you never actually show us what percentage of shops make it through your filters alive - and that number matters big time because if you're only keeping like 5% of scraped data, you're basically hemorrhaging API budget, but if you're sitting at 50%+, your scoring is probably way too generous. And your businessScore is looking at how recent posts are, but it's sleeping on engagement velocity - like how many likes and comments shops are actually racking up - which is usually a way better signal for finding shops that actually give a shit and will reply. If you threw together a quick A/B test showing that a businessScore ≥ 50 actually predicts higher reply rates in the real world, and now you've got a defensible case study instead of just a solid tutorial