Review data is worth exactly as much as it is fresh: a negative review from two weeks ago is no longer an event, it is a statistic. A chain with hundreds of stores cannot keep that freshness by hand. Scheduled review sync pulls new reviews from every connected source on a set cadence, sends them to analysis and raises alerts where needed. This post explains how sync is scheduled in BranchVoice, what is fetched per platform, and how an operations team keeps an eye on sync health.
Why scheduled review sync fetches only what is new
Re-downloading a store's entire review history on every sync is wasteful and expensive; pulling data from map and complaint platforms carries a cost per record. The guiding principle of the sync design is therefore incremental fetching: the first import collects history, and every later run picks up only what was added since the previous sync.
The principle is applied differently per platform:
- Google Maps: the last sync date is sent as the start date, so only reviews published after it come back.
- Yandex Maps: the platform offers no date filter, so periodic runs fetch only the newest reviews; anything already stored is not written again.
- Şikayetvar (Turkey's main consumer complaint site): after the first pull, a monitoring mode collects only complaints opened since the previous sync.
- Ekşi Sözlük (a large Turkish community forum): later runs are limited to the newest entries of each topic.
- X (Twitter): the review and lead passes run on separate schedules, each from its own last sync date.
A platform-level identifier prevents the same review being stored twice, so your counts do not inflate even if a sync runs again.
Cadence and timing
Every source has a sync frequency, chosen on the review settings step of the import wizard. Scheduled sync runs at most weekly; depending on your plan the options are weekly, monthly or "one-off pull only". More frequent automatic runs are deliberately not offered, because on platforms that bill the same records again they simply raise the cost; manual sync, by contrast, is always available.
Runs are scheduled at night: weekly sync on Sunday night, monthly sync on the first night of the month. They do not compete with daytime dashboard use, and the regional manager starts Monday morning with current data. The scheduler checks every hour for sources whose turn has come and queues them.
When to use manual sync
Three situations come up regularly:
- A crisis: when word of a complaint wave at a store reaches you, its map sources (Google and Yandex, if connected) are queued with one click instead of waiting for the next weekly run.
- A newly connected source: when a Şikayetvar or Ekşi Sözlük source is added, the first pull starts automatically; later it can be triggered from the scheduler list for an extra check.
- Before a report: ahead of the monthly management report, a regional manager refreshes their own stores by hand so the report goes out on the freshest data.
Manual triggering is done per source on the Sync Schedulers page, where each source shows the status and time of its last run.
What happens after a sync
A sync moves more than data; it starts a chain reaction. When new reviews are inserted:
- They are queued for AI analysis: sentiment, risk, topics and store matching.
- Alert rules are re-evaluated after a short delay; if a threshold is crossed, critical review alerts open and e-mails go out.
- When the analysis batch completes, the anomaly check is triggered.
In other words, nothing between "sync finished" and "regional manager received an e-mail" requires a person.
Monitoring sync health
The Review Queue page is the control tower for sync. At the top are counts of active, waiting, completed-in-24-hours and failed runs; below them a live list and a paginated history. Clicking a run opens a detail panel: which source, how many reviews were fetched, what kind of error occurred if any (platform access, timeout, quota) and the list of reviews that arrived with that run.
A failing source is not retried every hour; the wait grows with each failure, up to a week at most, so a broken source cannot drain the budget. Sources that keep failing, or that have delivered nothing for longer than twice their expected interval, appear in a data-health banner at the top of the dashboard and as a separate line in the weekly digest e-mail.
| Status | What it means | What to do |
|---|---|---|
| Failed | The last run errored and will be retried | Check the error type; for access errors, verify the source link |
| Stale | No new data for a long time | Trigger a manual sync and confirm the source is still active |
| Delayed | Waiting for a retry | Just wait; if it keeps happening, review the source settings |
For how sources are connected and matched to stores, see importing your store list.
Next steps
Scheduled sync is the set-and-forget part of the platform: you choose a cadence once, the rest runs overnight and shows up as alerts in the morning. To see the sync flow and the queue screen for your own sources, request a demo.