Tokens, Cost, Response Performance, and Devices
OpenAI Usage & Costs calls organization endpoints with the optional Admin API key and Project ID. Response duration and device/channel information come from local logs.
Token and cost cards explain OpenAI account consumption; response-performance and device views explain what SmartSite recorded locally. Input tokens include instructions, conversation and supplied context, while output tokens reflect generated text. Duration can also include routing, retrieval, tool, and network time, so a slow turn is not automatically a slow model.
Start with the smallest safe step: choose day, month, or year and the corresponding period. Do not consider the task finished before you measure the same period/test set again; this is where the configuration is tested in the context that truly consumes it.
What this feature does and when to use it
Section titled “What this feature does and when to use it”Use trends to find unusual spend, high token usage, slow responses, or traffic changes.
Use this feature in the following situations:
- You are monitoring spend or request volume for a chosen day, month, or year.
- You want to identify high-token or slow conversations and inspect their detailed causes.
- You are evaluating a model, instruction, or answer-length change with a repeatable test set.
Where to find it
Section titled “Where to find it”Before you begin
Section titled “Before you begin”- SmartSite Assistant is active.
- Local logs exist for performance/device metrics.
- An OpenAI Admin API key is configured for organization usage/cost cards; a regular key is not sufficient.
Set it up step by step
Section titled “Set it up step by step”- Choose Day, Month, or Year and the corresponding period.
- Refresh and note whether the result is live or cached/stale.
- Review total cost, request count, input tokens, and output tokens.
- Compare local response duration and device/channel mix over an equivalent window where possible.
- Open outlier conversations for model, token, routing, knowledge, and tool evidence.
- Change one cost/performance driver.
- Measure the same period/test set again.
Fields, controls, and important values
Section titled “Fields, controls, and important values”Cost, token, speed, and device data help balance answer quality with operating constraints. A more capable model may cost more, but high token use can also come from long instructions, oversized tool results, or unnecessary context. Performance information does not tune the AI automatically; it tells you what to test before changing models or shortening useful content.
| Field, control, or status | What SmartSite Assistant does with it | How to use it and why it matters |
|---|---|---|
| Admin API Key | Required for organization usage and cost endpoints. | Add “Admin API Key” when the organization-level cost charts are genuinely needed. It will not improve a single answer, and it should receive tighter handling than the normal runtime credential because of its wider reporting scope. |
| Project ID | Optional analytics filter. | Enter “Project ID” exactly as its source system provides it. This reference points the workflow at a particular resource; it does not improve AI reasoning, and the wrong value can quietly target the wrong account or stored record. |
| Cache | Per-period result cached for one hour; background refresh can retain stale data when live refresh fails. | Recent reporting data may be reused for an hour so the dashboard does not repeatedly call organization endpoints. A stale chart changes what the administrator sees, not what visitors are charged or how their responses are generated. |
| Total cost / requests / tokens | OpenAI organization data. | Read these organization totals as operating evidence, then compare them with model choice, instructions, conversation length, and tool results. The figures show consumption but cannot identify which content made an individual answer useful. |
| Response performance / devices | Computed from local SmartSite log rows. | Combine local duration and device information with individual conversation details to find patterns such as mobile-only friction or unusually slow tool calls. These measurements suggest what to investigate; they do not identify the cause or improve future responses without a tested change. |
How to confirm it is working
Section titled “How to confirm it is working”Use a separate test session to confirm Tokens, Cost, Response Performance, and Devices. This keeps existing login, browser storage, and response history from hiding the change, and it shows whether the result reaches the complete workflow rather than stopping at WordPress storage.
Practical example
Section titled “Practical example”Investigate rising output tokens by filtering high-output turns and checking whether instructions or retrieved context produce unnecessarily long answers.
Recommended practice
Section titled “Recommended practice”- Change one part of Tokens, Cost, Response Performance, and Devices at a time and keep a short record of the previous value and test result.
- Turn one observed pattern into one controlled configuration change, then compare new conversations with the earlier evidence.
Important warnings
Section titled “Important warnings”Common problems and focused checks
Section titled “Common problems and focused checks”| Problem | What to check and what to do next |
|---|---|
| Tokens, Cost, Response Performance, and Devices is missing or does not match this guide. | Confirm the plugin is active and the account can manage WordPress options. Narrow the date and filter scope, then open source records instead of relying on an empty chart or summary alone. |
| A change on Tokens, Cost, Response Performance, and Devices does not produce the expected result. | Keep the exact notice and test case, then review the browser console and WordPress/PHP log. Narrow the date and filter scope, then open source records instead of relying on an empty chart or summary alone. |
Screen reference
Section titled “Screen reference”- Capture
- Show a Month analytics view with sanitized cost/request/token cards, response chart, and device chart; annotate data-source boundaries.
- Show
- Period controls, cache status, cost/request/token cards, performance and device charts
- Viewport
- Desktop, 1440 × 900
- Annotate
- Use numbered callouts only for controls referenced in the procedure.
- Redact
- OpenAI keys, tokens, secrets, personal information, private URLs, IP addresses, and conversation text