Best practices
The habits that separate clean, cheap, reliable workflows from credit burners. Most of these come straight from the team that builds workflows with customers every day; the video version is Floqer 101, Best Practices.
1. Test on 20 rows, not 2,000
Always run a new or edited workflow on a small sample first: 3 to 5 rows to check correctness, 20 or so to check quality. Per-row provider calls cost credits, and mistakes are cheaper at small scale. Only run wide once the sample comes back verified.
2. Budget for R&D
Dedicate 15 to 20% of your credits purely to experimentation and testing. Treating that spend as a planned budget, instead of an accident, is what lets you iterate on providers, prompts, and qualification logic without anxiety.
3. Order waterfall providers best-first
In a waterfall, provider order is your main cost lever. Put the providers with the highest hit rate and lowest cost first, and turn verification on so only good data flows downstream. Mix premium and budget providers to find your sweet spot.
4. Use the free building blocks
Data formatting, cleanup, Merge Lists, and the other built-in tools cost nothing. Zero-credit fields and the AI prompt generator help you shape data and prompts without spending. Reach for the free tools before the paid ones.
5. Bring your own API keys
If you already pay for Apollo, Hunter, Prospeo, or similar providers, connect your own API keys. Your existing contracts then cover those calls, which can cut workflow costs substantially.
6. Qualify before you enrich
Use conditional runs so expensive enrichment only happens on rows that pass your qualification checks. Enrich strategically, not everything: start with qualification signals (headcount, tech stack, job postings), then layer personalization data (posts, reviews, job changes) on the rows that survive.
7. Let the cache save you money
If the same information was already enriched on Floqer before, you get it for free. Prefer common, cacheable lookups where quality is equal.
8. Design for the person waiting
If a workflow backs a shortcut that reps run live, speed matters: fewer providers in each waterfall means a faster run. Put slow endpoint work (CRM writes, sequence enrollment) after the output node, so the rep sees the result while the rest completes in the background.