Beyond day 90: maintenance cadence

A high score is harder to hold than to reach. Competitors are also seeding citations. AI training data refreshes. New monitor prompts get added. Without a maintenance rhythm, a 75 in May is a 60 by August.

The daily rhythm (10–15 min)

• Open /dashboard/proposals. Approve, reject, queue. Most days this is one or two clicks.
• Review customer-visible Activity and Logs for incomplete work. Contact support about persistent errors.
• Execute the day's tactic. One thing, done well.

That's it. If you can't sustain 10–15 min/day, the maintenance cadence isn't real.

The weekly rhythm (45 min, one session)

Pick a day, Friday is good, and block 45 minutes:

1. Read the week's monitor scan results. Note any SoM drops >5%. Diagnose: did a competitor land a press hit? Did your most-cited Reddit thread get archived? Address.
2. Run a citation batch. 10–15 new placements. Keep the volume steady; lumpy citation cadence creates lumpy SoM data.
3. Send 2–3 press pitches. Press is the pillar that decays slowest, but only if you keep feeding it.
4. Record what went live. Add implementation evidence for approved work in the Work Queue.
5. Check unanswered alerts. If you've got more than five backed up, you've got too many rules.

The monthly rhythm (2 hours, plus reports)

At the start of each month, use the same week to:

1. Generate the report. Open Reports, generate last month's report, verify the figures and share it when it is ready. Automatic delivery on the 1st is off unless Outrigger enables it.
2. Run a fresh audit. Compare to last month. If a pillar dropped, dig in.
3. Re-evaluate competitors. Add anyone new the platform has discovered. Drop anyone who's no longer relevant.
4. Refresh keyword research. Run a new seed. Promote one or two new clusters.
5. Review credit usage. Open Billing and credits, then check activity by brand for unexpected usage.

The quarterly rhythm (half a day)

Once a quarter, do a strategic pass:

• Re-read the brand brief. Has the business changed? Update accordingly. Every downstream tactic depends on it being current.
• Re-evaluate the monitor prompt set. Are these still the buying-intent questions humans ask AI? Or has the category vocabulary shifted?
• Check the benchmark page. Has your bucket's median moved? Are you still p75+?
• Decide: do you want to push for p90+ or hold p75? Pushing higher costs disproportionately more press budget.

What decays without maintenance

The fastest pillars to lose:

• Citations: threads get archived, AI training data refreshes. Without 10+ new placements per week, SoM gradually drifts down.
• Reviews: velocity matters. Three months of zero reviews and the freshness signal weakens.
• Press: old coverage stays old. Without new mentions, AI models read the brand as static.

The most durable pillars:

• Entities: once fixed, mostly stays fixed. Re-audit quarterly to catch drift (a directory updates its UI, your description gets truncated, etc.).
• Wikipedia: once you have a stub, it's hard to lose unless deleted by an editor.

So the maintenance budget is: spend most of the time on citations, press, and reviews. Light-touch on entities. Defend Wikipedia with a watchlist.