Insights · Reliability

Uptime is a reputation problem once you have many locations

When a company runs dozens of sites, profiles, and phone lines, small failures stop being IT issues and start being brand issues. Here is how to watch for them without drowning in alerts.

Most leadership teams think of uptime as an engineering metric. It lives on a dashboard somewhere, it is usually green, and it only becomes a conversation when something large goes down. That view works for a single website. It breaks down once a company operates many locations, brands, or regional sites.

At that scale, the failures customers actually experience are rarely dramatic. They are small and local: a location page that returns an error after a site update, a booking link that points to a retired system, a security certificate on a regional domain that quietly expired, a map listing that shows the wrong hours because someone suggested an edit. None of these trip a company-wide alarm. All of them send a customer somewhere else.

Why the small failures matter more than the big ones

A major outage is visible, so it gets fixed fast. Everyone notices, someone owns it, and the incident ends. Small failures are the opposite. They are invisible to headquarters, they often sit for weeks, and the people who notice them first are customers who do not report them. They simply call a competitor.

The damage also compounds in places you do not control. A customer who hits a dead page may leave a frustrated review. A wrong phone number on a listing sends calls to a closed line, which looks to the caller like a business that does not care. Search engines and map platforms treat broken pages and inconsistent details as signals of quality. Over time, a pile of small failures becomes a reputation, and reputations are slow to repair.

What to watch, and how often

The goal is not to monitor everything. It is to monitor the few things that sit directly on the path between a customer and a conversation with you. For most multi-location companies that list is short:

  • Reachability of every customer-facing domain and key page. The home page, each location page, contact and booking paths. Check frequently, and alert only after repeated failures so a single blip does not wake anyone up.
  • Certificate expiry on every domain, including the forgotten ones. Regional and campaign domains are where certificates lapse, because nobody remembers they exist.
  • The things a customer clicks. Phone links, email links, buttons, and forms. A page can be up while its main button is broken. Check these on a weekly crawl, without submitting real forms.
  • Public listing details for each location. Name, address, phone, hours, and open or closed status. These change more often than most teams expect, sometimes because of well-meaning outsiders.
  • Phone-sized layout. Many customers arrive on a phone. A layout that breaks on a small screen is functionally an outage for them.

Ownership is the real fix

Monitoring only helps if every alert has an owner. In multi-location companies, ownership is often the gap. The web team owns the platform, marketing owns the listings, operations owns the locations, and an agency owns a few microsites from a past campaign. When a location page breaks, each group reasonably assumes it belongs to someone else.

A simple ownership map solves most of this. For every domain, page type, and listing, write down who fixes it and who is told. Keep it to one page. Review it whenever a location opens, closes, or changes hands, and after any acquisition, because acquired brands arrive with their own domains, accounts, and habits.

Make alerts calm and specific

Alert fatigue is the most common reason monitoring programs fail. If people receive constant noise, they stop reading, and the one alert that matters is lost. A few rules keep alerts useful:

  • Require repeated failures before alerting, and always send a recovery notice so people know the issue is closed.
  • Warn once, well ahead of a certificate expiring, rather than every day.
  • Say what broke, where, and who owns it, in plain language. An alert should be actionable by someone who is not an engineer.
  • Roll the rest into a short weekly report rather than a stream of messages.

Report it like a business metric

Leadership does not need raw check logs. It needs to know whether the customer path is healthy across the business, where it is not, and whether that is getting better. A monthly one-page view works well: which locations had issues, how long they lasted, what was fixed, and what is still open. Compare each month against a baseline taken before the program started, so progress is measured rather than assumed.

Handled this way, uptime stops being a background technical number and becomes what it really is at scale: a running measure of whether customers can reach you, everywhere you operate.

A starting checklist

  • List every customer-facing domain, including old campaign and regional sites.
  • Assign an owner for each domain, page type, and location listing.
  • Set reachability and certificate checks with repeat-failure alerting.
  • Run a weekly crawl of links, buttons, and forms, read-only.
  • Take a baseline now, and report against it monthly.

More insights