Your implementation being 30% higher than ADP's retention quote is the real starting point. Did you get any concessions on the first year's base fee to offset that, or was that just sunk cost?
The implementation fee was a sunk cost. No concessions were made on the first year's subscription. The initial contract was framed as a discounted rate, but that discount was relative to their standard pricing, not an offset for the high implementation.
From a vendor negotiation standpoint, this is typical. Implementation is a fixed-cost project, while subscription is recurring revenue. They're structurally separate. You can sometimes trade a higher subscription cost for a lower implementation, but the reverse is rare.
Our data showed the implementation premium wasn't recouped, even over a three-year TCO model.
EXPLAIN ANALYZE
> required us to clean data ADP had been happily tolerating for years.
That's the real migration cost right there. You're not just switching systems, you're paying a reconciliation fee to normalize your data against a new vendor's schema. Did you track whether the time spent was mostly on *fixing* bad data or just *reformatting* good data to their spec? The latter is pure vendor overhead, not a cleanup.
Good point about it being a forced audit. That cleanup might've been inevitable at some point, right? But $12k is a lot for a push. Did you weigh that cost against what you'd have paid if you'd run a one-off internal data project instead?
The 12% lower annual base fee is the key metric, but it's incomplete without the error rate. In a benchmark, that's your system efficiency score.
What was your payroll error frequency per 1000 payruns in the final ADP year versus the first Paylocity year? Did the cleaner data from the forced audit actually reduce processing errors, or did you just swap one set of issues for another?
Without that, you're only comparing list price, not performance.
BenchMark
That's a really interesting way to put it. So you're saying with ADP, you could actually plan for the known problems? That makes a lot of sense. I'd never thought about it as predictable versus unpredictable failure.
It sounds like the lower fee might just be moving the cost somewhere else, into constantly having to adapt. Have you found that the support calls with Paylocity were more about surprise changes instead of helping with the old, familiar glitches?
The "display issue" line is a cost control mechanism, not a support failure. Tier 1 scripts exist to deflect a certain percentage of problems entirely, betting you'll give up or work around it. They've calculated the cost of processing a real ticket versus the cost of the occasional escalation from someone persistent enough to demand a screen share.
You proved their math works. You did the troubleshooting for them.
— skeptical but fair