The in-house build threat requires careful calibration. In my integration work, I've seen it backfire when the vendor's technical team recognizes the architectural bluff. They'll call it by asking specific questions about your planned stack's scalability or compliance features.
A more effective approach I've used is to genuinely map the data flows you'd need to replicate. Document the API call volume, the transformation logic, and the monitoring hooks. Presenting that mapping exercise - which is real work you'd do anyway for integration - lends credibility. It shows you understand the cost components of their service, not just the sticker price.
This shifts the conversation from a vague threat to a discussion about their platform's unique value. You might not get a discount, but you often uncover bundled services or future roadmap items that change the cost-benefit analysis entirely.
Spot on about using their own pricing as leverage. I've done that with marketing automation platforms by comparing their per-contact pricing for monthly plans to their annual enterprise tiers. It forces the conversation away from market comps they can dispute.
That said, the time-cost caveat is huge. I'd push your $10k threshold even higher for niche tools. If it's truly a must-have and the evaluation already took months, a last-minute discount chase on a $15k deal can sour the implementation vibe. The savings just aren't worth the relationship tax.
Always optimizing.