Workday OfficeConnect Refresh Time Benchmarks
“Is my workbook slow?” is a question with no universal answer — refresh time depends on formula count, server load, network path, and how complex your Adaptive Planning model is. This reference gives you rough Workday OfficeConnect benchmarks to compare against, and a method for measuring your own.
What you’ll need:
- A Workday OfficeConnect workbook you can refresh end-to-end
- A stopwatch (your phone is fine) or the Excel status bar
Benchmark table
The numbers below assume a healthy Adaptive Planning tenant, a wired or strong wireless network, and 64-bit Excel. Cut the workbook into one of these buckets by counting OfficeConnect formulas (Reporting pane → Review tab → Workbook Elements count).
| OfficeConnect formulas | Expected refresh | What “slow” looks like |
|---|---|---|
| Under 50 | Under 5 seconds | Over 10 seconds |
| 50–200 | 5–10 seconds | Over 20 seconds |
| 200–500 | 10–30 seconds | Over 60 seconds |
| 500–1,000 | 30–60 seconds | Over 2 minutes |
| 1,000+ | 1+ minute | Over 5 minutes |
A workbook in the “slow” column isn’t necessarily broken — it might be running against a heavily-loaded tenant or pulling unusually complex modeled accounts — but it’s almost always optimizable.
Factors that move the benchmark
Even within the right formula bucket, several variables shift expected refresh time:
- Tenant load. Peak budget season (typically Q4 and Q1) and end-of-month close can double or triple response times. The same workbook can refresh in 8 seconds at 7 a.m. and 25 seconds at 2 p.m.
- Network path. A corporate VPN routing through a distant data center adds latency to every server call. A 100-formula workbook on direct internet may take 6 seconds; the same workbook over VPN can take 20.
- Model complexity. Modeled (calculated) accounts cascade through formulas server-side before returning a value. A report pulling 50 modeled accounts is meaningfully slower than 50 input accounts.
- Time granularity. Monthly columns make 12x the server calls of annual columns covering the same year, all else equal.
- Level scope. Pulling all 200 cost centers is slower than pulling one division, even when the cells aggregate to the same number.
How to measure your own refresh time
When to take action
Apply the rule of thumb: if your measured time is more than 2x the benchmark for your formula bucket, optimize. If it’s within the bucket, accept it — further tuning has diminishing returns and risks breaking the report.
If it’s far outside the bucket (10x+) and you’ve ruled out network and tenant load, something specific is wrong: an oversized hidden range, a runaway repeating row, or a modeled-account explosion. Use Cell Explorer to find which cells are taking the longest.
Result
You have a defensible answer to “is this workbook slow?” — measured against a benchmark, not a feeling — and a triage path for what to do next.
Next steps
- Optimize Performance for Large Models — the canonical tuning checklist once you’ve decided action is warranted.
- Reduce OfficeConnect Element Count — the single biggest lever, in detail.
- Fix Slow Performance in Large Reports — diagnostic walkthrough when refresh is far outside the benchmark.