Every review, every platform, one inbox

Why reviews are the cheapest lever nobody pulls

Ask a local business owner what most influences whether a stranger rings them, and if they have been paying attention they will say reviews. Ask them when they last replied to one, and the answer is usually a shrug. The gap between how much reviews matter and how much attention they receive is the widest in local marketing, and it is almost entirely a logistics problem rather than a belief problem.

The logistics are genuinely awful. A typical trades business is reviewed on Google, on Facebook, on one or two trade directories, and possibly on Trustpilot or Yelp. Each has its own login, owned by whoever set it up, which is frequently a former employee or the owner's son-in-law. Nobody logs in regularly. New reviews are discovered by accident. A furious one-star can sit at the top of the profile for six weeks before anyone notices, during which every prospect who searched the business name has read it.

For an agency the problem multiplies. Twenty clients across four platforms is eighty logins, and no agency in the world is checking eighty profiles weekly. So reviews become the thing that is mentioned in the pitch, agreed to be important, and then never systematically managed — which is a shame, because they move local rankings, they move conversion rate on every page of the site, and they cost nothing but attention.

The effect on ranking is the part agencies undersell. Review count, rating and recency are among the strongest signals in the local pack, and unlike most ranking factors they are almost entirely within the client's control. A business that goes from eleven reviews to sixty over a year, with replies on all of them, will usually see local visibility move more than a year of technical SEO would have achieved on the same site.

The effect on conversion is even larger and easier to demonstrate. The same landing page, the same ad, the same budget — with a 4.8 from 200 reviews rather than a 3.9 from 12 — converts substantially better, because the prospect checks. Every agency running paid traffic to a business with a weak review profile is paying for clicks that a stronger profile would have converted.

What the inbox actually does

The scope is deliberately narrow: see everything, respond quickly, know when something needs attention, and be able to report on it.

Getting a client set up

The hard part is not technical. It is getting access to profiles the client may not realise they own.

1. 1

   Connect what can be connected

   Google Business Profile and Facebook authorise through OAuth, which takes a minute and gives full read and reply access. These two are most of the value on most accounts.

2. 2

   Add the platforms that only allow reading

   Some directories expose reviews but not replies, or expose neither through an API. Those are monitored where possible and clearly marked as read-only, so nobody writes a reply that cannot be delivered.

3. 3

   Claim what has not been claimed

   A surprising number of local businesses have profiles they have never claimed, accumulating reviews they have never seen. Finding those is often the single most valuable hour of an onboarding.

4. 4

   Decide who replies and how fast

   Agency replies on the client's behalf, client replies with agency drafting, or agency handles negatives only. All three work; what does not work is leaving it undecided, which reliably means nobody replies.

5. 5

   Set the alert threshold

   Usually anything three stars or below, to whoever can act at the client. An alert that goes to an account manager who then emails the client has added a day to the response time.

Replying well, which is harder than replying fast

Response speed is necessary and not sufficient. A fast bad reply is worse than a slow good one, and the failure modes are predictable enough to be worth naming.

The defensive reply is the most common and the most damaging. A one-star complaint gets a response explaining why the customer is mistaken, listing what the business did correctly, and ending with a note that the customer failed to follow instructions. Every word of it may be true. It reads, to the two hundred prospects who see it, as a business that argues with unhappy customers, and it does more damage than the original review would have done unanswered. The reply is not addressed to the reviewer; it is addressed to everyone reading later, and that is the audience to write for.

The template reply is the second failure. Six identical "Thank you for your feedback, we value your business" responses in a row are visible to anybody scrolling and communicate that nobody actually read any of them. Templates should be a skeleton somebody fills in with a detail from the specific review — the job, the location, the technician's name. Thirty seconds of personalisation is the difference between a reply that helps and one that mildly hurts.

The third is ignoring positives. Most businesses that reply at all reply only to complaints, which produces a profile where every response is defensive. Replying to good reviews is quick, it demonstrates engagement to future readers, and on some platforms it correlates with visibility. It is also the easiest thing in this entire category to delegate.

The fourth, rarer but worth naming, is the reply that leaks. Someone responding on behalf of a clinic, a law firm or any business with a duty of confidentiality can confirm a person was a customer, or reference their circumstances, simply by answering the review at all. In regulated sectors the safe reply says nothing about the individual and offers a channel — and in some of them, the safe reply is drafted by the client rather than the agency.

The pattern that works, in three parts: acknowledge the specific thing, take responsibility for whatever part is genuinely yours without over-apologising, and move the detail off the platform with a name and a direct contact. Short. Never argue with facts in public, even when you would win.

What each platform actually allows

Capability varies enormously and vendors are frequently vague about it. Being explicit prevents an awkward moment when a client asks why their Yelp replies are not appearing.

Asking for reviews without getting the client into trouble

Volume and recency matter, which means asking. It also means asking carefully, because this is the part of reputation management with real rules attached and real consequences for breaking them.

The line worth understanding is gating. Sending every customer a request is fine. Sending a survey first and only inviting the happy ones to review publicly is review gating, it is against Google's policies, and profiles have been penalised for it. It is also extremely common in reputation software, often as a headline feature, described in language designed not to draw attention to what it is. An agency that deploys it across an account base is putting its clients' primary marketing asset at risk to improve a number on a report.

Incentives are the other trap. Offering a discount for a review breaches the terms of most platforms and, in several jurisdictions, consumer protection law. It is also transparently detectable in the review text, which tends to produce exactly the pattern-matched cluster that gets a profile flagged.

What works is unglamorous. Ask everyone, ask soon after the work is done, ask in the channel the customer already uses — usually a text message — and make the link one tap. Response rates on a well-timed SMS request are several times an email's. A business that asks every customer for six months will accumulate more reviews than one running any clever scheme, and will still have a profile in twelve months.

The honest note for agencies: this platform does not send review request campaigns. It monitors, replies and reports. Where a client needs a request workflow, that is typically their CRM or a dedicated tool, and it is better to say so than to ship a half-built sender that gets somebody penalised.

What makes this workable at agency scale

Twenty clients across four platforms is the case that matters. Everything here is built for that rather than for one business watching its own profile.

• Cross-client triage. One view of every unanswered review across every account, sorted by urgency. An account manager's Monday is a list of eleven items rather than twenty dashboards.
• Alerts go to whoever can act. Routing is per client, so a negative review can reach the client directly, the account manager, or both, without a relay step that costs a day.
• Replies are attributed. Who wrote and sent each reply is recorded, which matters when a client queries the tone of something posted in their name.
• Access is scoped per client. A client login sees their own reviews only, enforced server-side. Nobody reaches another client's reputation data by changing a number in a URL.
• Connections are monitored. OAuth tokens expire and get revoked. A platform that has stopped syncing is reported as disconnected rather than quietly showing no new reviews, which looks identical to a quiet month.
• It is white-labelled. Reputation reporting carries your branding on your domain, on plans that include white label.

Selling reputation work without selling a promise you cannot keep

Reputation management has a reputation problem of its own, and agencies pitching it inherit the scepticism. Enough businesses have been sold review removal, review suppression or "guaranteed five-star profiles" that a client hearing the phrase assumes something slightly disreputable is being offered. Distinguishing yourself from that is mostly a matter of being precise about what you will do.

You cannot get reviews removed. Platforms remove reviews that breach their own policies — reviews from competitors, reviews containing abuse, reviews for a business the person never used — and flagging those is worth doing, but success rates are modest and unpredictable. A genuine unhappy customer's honest one-star is going to stay, and any vendor implying otherwise is either failing or doing something that will eventually get the profile suspended.

What you can promise is process. Every review will be seen. Negatives will be answered within a stated window. Positives will be acknowledged. The client will know what their profile says about them without having to check. And the volume and rating trend will be reported monthly beside the lead numbers. That is a service level rather than an outcome, which is exactly right, because the outcome depends on the client's work quality and nobody should pretend otherwise.

The interesting consequence of framing it that way is that it makes the awkward conversation possible. When a client's rating is genuinely poor because their service is genuinely inconsistent, no amount of reply-writing fixes it, and the agency that says so — with the reviews in front of them, quoting the same complaint appearing eleven times — is providing far more value than the one that quietly keeps writing polite responses. That conversation has saved accounts. It has also, occasionally, ended them, which is a cost worth accepting for the ones it saves.

Reporting reputation next to leads

Reputation reporting on its own is thin. Rating went from 4.4 to 4.6 and volume rose by thirty-one; the client nods and nothing follows. It becomes interesting when it sits beside the lead data, because then the question is whether the reputation work produced business.

Sometimes the relationship is visible. A client whose review volume tripled over a year and whose organic and map-driven calls rose over the same period has a plausible story, and while it is correlation rather than proof, it is considerably more than most reputation reporting offers. The absence of a relationship is also informative — a client whose rating improved substantially with no movement in enquiries may have a visibility problem rather than a trust problem, which is a different piece of work.

There is one number worth putting on the report that almost nobody does: response rate and median response time. It is the only part of the reputation picture entirely within your control, it is directly attributable to the work you were paid for, and it is the metric a client can check independently by looking at their own profile. Reporting "94% of reviews answered, median response 3 hours" alongside the rating trend separates the service you delivered from the outcome you influenced, which is exactly the distinction that keeps a reputation retainer from being judged on something you do not control.

The practical framing for a client meeting is not "your rating improved". It is "you now have sixty reviews at 4.8 rather than eleven at 4.1, you appear in the map pack for these terms where you did not, and calls from that source are up this much". That is one narrative built from three parts of the same system, and it is the kind of thing that gets a retainer renewed without a negotiation.

Common questions