Everyone's competing online. We make sure you're the one they find.

If you run a clinic, a showroom or a service business, the map pack sits above every organic result. Getting into it is less about your website than most agencies admit — it is about the profile itself, and about proximity, relevance and prominence.

Proximity you cannot change. Relevance and prominence you can. Relevance comes from your primary category being exactly right — "Dental clinic" not "Doctor", "Orthodontist" as a secondary rather than a primary if that is not your main revenue. We have moved clinics several positions by fixing nothing but the category.

Prominence comes from review velocity and photo freshness. A profile that gains four reviews a month steadily outperforms one that got forty in a burst two years ago. Ask at the point of maximum goodwill, which for a clinic is right after treatment finishes, not in a follow-up SMS three days later.

The unglamorous part: fill in your services and products individually rather than leaving them to be inferred, keep your hours accurate including holidays, and post something once a fortnight. None of that is clever. It is just the work almost nobody does consistently.

When a client tells us analytics looks healthy but nothing is coming in, we look at four things before touching design.

First, load time on mobile data. Not on your office wifi — on a mid-range Android on a patchy 4G connection. If the page takes more than four seconds to become useful, a third of visitors are gone before they see anything you wrote.

Second, whether the next step is visible without scrolling. A phone number in the footer is not a call to action. On a phone, a tap-to-call button should be reachable from anywhere on the page.

Third, whether the page answers the question the visitor arrived with. Someone searching for a price does not want your company history. Match the page to the intent that brought them.

Fourth, trust signals near the action, not in a separate section. Reviews, credentials and real photographs belong beside the button, because that is the moment doubt appears.

Fix those four and the same traffic converts differently. We have seen enquiry volume double with no change to visitor numbers at all.

We audit a lot of Meta accounts before taking them over. The same three problems appear again and again.

Boosting instead of campaigns. The boost button optimises for engagement, because engagement is what the button was built for. Likes are not enquiries. Running the same creative through Ads Manager with a lead or conversion objective typically costs less per actual result, sometimes dramatically less.

Audiences squeezed too tight. Layering six interests and three behaviours on top of a small city radius leaves the algorithm almost nothing to work with. Modern delivery finds buyers better than manual stacking does. Give it a broad audience, a clear conversion signal, and strong creative, and get out of the way.

Judging on the wrong number. Cost per click flatters campaigns that bring the wrong people. Cost per qualified enquiry is the only figure that matters, and it means tracking what happens after the click — which means your pixel and your lead form need to talk to whatever you use to record enquiries.

Creative volume beats creative perfection. Four honest variations tested for a week will teach you more than one polished ad running for a month.

Agencies love to quote a starting budget. The number is usually chosen to make the retainer look reasonable rather than to reflect your market.

Work backwards instead. Take the average value of one customer to you over a year. Take the share of enquiries you actually convert. Multiply, and you have the most you can afford to pay for an enquiry while still making money.

A clinic where a new patient is worth twenty thousand rupees over a year, converting one enquiry in three, can afford well over a thousand rupees per enquiry and still be comfortably ahead. A business selling a two-thousand-rupee product once cannot.

Now check that against what clicks cost in your category. If a click costs sixty rupees and one enquiry takes fifteen clicks, an enquiry costs nine hundred. Compare with your ceiling. If the gap is thin, ads are not your first move — your website or your conversion rate is.

Only after that does a monthly figure make sense, and it should be large enough to gather meaningful data within a few weeks. Splitting a small budget across search, display and shopping simultaneously guarantees you learn nothing about any of them.

An enquiry arrives on WhatsApp at eight in the evening. Whoever is on the phone is busy. It gets seen the next afternoon. By then the person has messaged two other places and booked with whoever replied first.

This is not a discipline problem. It is a systems problem, and it is fixable without hiring anyone.

The minimum useful setup: every enquiry, from every channel, lands in one place. An automatic acknowledgement goes out within a minute, saying when a human will reply. Anything unanswered after a set period escalates to somebody by name.

That alone recovers a surprising share of lost enquiries, because most people are not comparing carefully — they are choosing whoever responded while the intent was still fresh.

The second layer is memory. Someone who enquired about implants in March and did not book should not get a generic message in June. A short, specific follow-up referencing what they originally asked about converts far better, and it only works if the original detail was captured automatically rather than remembered.

None of this requires a large platform. It requires the pieces you already use to be connected.

The failure mode of business chatbots is ambition. Given a wide brief, they answer questions they should not, invent details, and leave the customer more annoyed than before.

Give it the boring work instead. Opening hours, location and directions, what a first appointment involves, what to bring, whether you take a particular insurance, roughly what a common service costs. These are asked constantly, the answers rarely change, and getting them instantly at eleven at night is genuinely useful.

Draw a hard line around anything clinical, financial or contractual. A bot should never assess symptoms, never confirm what a treatment will cost in a specific case, never promise an outcome. It should recognise those questions and hand over cleanly: acknowledge the question, say a person will answer, capture the detail, and flag it.

The handover is the part that decides whether people trust it. If a bot loops instead of escalating, you have made your service worse. Every conversation needs an obvious route to a human, and that route needs to actually work outside office hours by taking a message rather than pretending.

Written that way, a bot is not a replacement for your team. It is a filter that means your team only sees the conversations worth their time.