I run Google Ads accounts for clients at Limi Creatives and delivered paid search alongside SEO for Minnesota-based US clients during a remote contract with SoloiTech from December 2025 to June 2026. When you hire me you get the person who logs into the account, not a strategist summarizing someone else's work on a monthly call.

What follows is what I do and what I think, including the parts that lose me work.

The conversion action most accounts are optimizing toward by mistake

Before structure, before keywords, before bidding, I look at one screen: the conversion actions table, and which of them are marked primary. This is where most accounts are quietly broken.

The pattern I see repeatedly is an account with eight or nine conversion actions all marked primary: contact page views, clicks on a phone link nobody dialed, a newsletter signup, a PDF download, a form submit. Smart Bidding treats every primary action as worth buying, so it buys the cheapest one. That is almost never the one that pays your rent.

The fix is boring and it works. One primary conversion action representing money or a qualified lead. Everything else demoted to secondary, where it still reports but no longer steers bidding. If calls and forms both genuinely matter, conversion values say how much more one is worth rather than pretending they are equal.

Then the counting setting. "Every" versus "one" conversion is not a detail on a lead-gen account. If a prospect submits your form three times because the confirmation page loads slowly, an account counting every conversion learns that person was worth three of the one who submitted once.

Then the conversion window. If your buyers take nine weeks to decide, the default thirty-day click window teaches the bidding algorithm that most of your good traffic was worthless.

Reading the search terms report like an invoice

The keyword list tells you what you asked for. The search terms report tells you what you bought. Those are different documents and only one of them is honest.

With broad match and Smart Bidding now the combination Google pushes hardest, that gap has widened. Broad match can work well, but only on a clean conversion signal with an actively maintained negative list. Broad match plus maximize conversions plus the wrong primary conversion action is the fastest way I know to spend a month's budget on traffic that was never going to buy.

I do a weekly pass through search terms in the first month, sorted by cost rather than clicks, because the expensive irrelevant term matters more than the frequent cheap one. Negatives go in at three levels: account-level for permanent exclusions such as "free", "jobs", "salary" and "DIY", campaign-level for intent mismatches, and ad group level for cross-serving between related groups.

The underestimated part is shared negative list hygiene. Most accounts I audit have two or three overlapping lists applied inconsistently, so a term blocked in one campaign is still being bought in another. That is entropy rather than strategy failure, and it costs money every week nobody cleans it.

Structure: fewer campaigns, sharper intent

Sprawling accounts are a legacy of an older Google Ads where tight single-keyword ad groups gave real control. With close variant matching and machine bidding that control is mostly gone, and the sprawl now does harm by splitting conversion data across too many places.

I rebuild around intent tiers rather than keyword themes. High-intent transactional terms get their own campaign and budget, because those can never afford to be throttled. Comparison and research terms sit separately, with a lower target and different landing pages, because they convert on a different timeline. Brand gets its own campaign so it never subsidizes the rest of the account and you can see what brand actually costs you.

Where I keep granularity is landing page alignment. An ad group exists because it points somewhere specific, not because someone wanted a tidier spreadsheet. If two ad groups send traffic to the same page with the same offer, they are one ad group.

Responsive search ads get pinned selectively. I pin the offer or qualifying term in headline position one when it prevents a misleading combination, and otherwise leave Google to assemble. Over-pinning turns an RSA into an expanded text ad and throws away the only thing the format is good at.

Performance Max, and the conditions under which I will run it

I use Performance Max when the situation earns it, and I will tell you when it does not.

It suits ecommerce with a real product feed, decent conversion volume, and an appetite for reach beyond search. Under those conditions I run it with brand exclusions applied, a separate brand search campaign keeping brand traffic out of it, search themes used sparingly, and asset groups split so the reporting means something.

It suits lead generation on a small budget much less well. A PMax campaign on a thin conversion signal will find the cheapest form fill on the internet and report it proudly. For most local service businesses I would rather put that money into a properly negatived search campaign where I can see the term behind every click.

I also reject the framing that PMax is a black box you simply trust. Asset group reporting, in-campaign search terms insights, brand exclusions and channel signals give more control than the default setup implies. Most accounts never touch them.

Getting closed revenue back into the account

The biggest single lever on a lead-gen account is not bidding. It is teaching Google which leads turned into customers.

If you close business over the phone or in a CRM, the account can only optimize toward form submissions unless you send the outcome back. I set up offline conversion tracking so the GCLID, or wbraid and gbraid for iOS traffic, is captured on the form, stored against the CRM record, and uploaded back to Google Ads when the deal is marked qualified or won. Enhanced conversions for leads is the alternative when the GCLID is hard to persist, matching on hashed email or phone instead.

This changes what the account aims at. Instead of buying the cheapest lead, it buys the lead most like the ones you closed. I would rather spend month one building that pipe than adjusting bid targets on a signal that was never measuring the right thing, which is why I own conversion tracking that reports the right numbers as part of the engagement rather than treating it as somebody else's dependency.

Three things I will not do in your account

I will not bid on competitors' brand terms as a default tactic. It occasionally works, more often it buys expensive clicks from people with no intention of switching, and it invites the same treatment back. If we do it, that is a deliberate test with a defined budget, not a line item I add quietly.

I will not run a display remarketing campaign purely to make the monthly report look busy. Display in a small account is usually where budget goes to produce impressions and view-through conversions that would have happened anyway.

I will not change bid strategies weekly. Every switch resets the learning period, and an account that is permanently in transition never produces a stable number to judge. If a strategy is going to be tested, it gets long enough to be judged fairly.

What the first month looks like

Week	What happens	What you see
1	Conversion action audit, tag and GA4 verification, search terms and negative list review, budget and bid strategy assessment	A written audit naming what is broken and what it is costing
2	Structure rebuild, negatives applied, ad copy and extension rewrite, landing page alignment flagged	The new account map, with the reasoning per campaign
3	Offline conversion or enhanced conversions setup where the sales process needs it, bid strategy set and left alone	Confirmation that closed business is flowing back into the account
4	First read on the restructured account, search terms pass, creative and asset iteration	A short report on what changed, what moved, and what is still noise

I do not promise a result inside thirty days. A restructured account needs a learning period, and anyone guaranteeing otherwise is selling you something. What I commit to is that by the end of week one you will know precisely what is wrong, whether or not you continue with me.

Whether I am the right person for this

Google Ads captures demand that already exists. If nobody is searching for what you sell, search will not manufacture that interest, and you would be better funded on paid social for demand generation rather than capture. I say this before invoices, not after.

If your category is searched but competitive, paid search is often the fastest way in while slower work compounds. Running it alongside organic rankings that reduce paid dependency over time is how a small business eventually stops renting all of its traffic. The full version of that trade-off, including timelines and cost, is in my breakdown on choosing between search and social spend.

You keep ownership of the account, tags, conversion actions and data. I work inside your account as a linked manager, and if we part ways nothing leaves with me. You can look at accounts I've restructured, and when you want a straight answer on what yours is doing wrong, get your account reviewed.

Frequently asked questions

Why is my Google Ads account getting cheap leads that never buy?

Almost always because bidding is optimizing toward the wrong conversion action. If page views, PDF downloads and click-to-call links are all marked primary, Smart Bidding buys the cheapest of them. Set one primary action that represents a qualified lead or revenue, demote the rest to secondary, and import closed deals so the account learns what a good lead looks like.

Should I use broad match keywords in Google Ads?

Broad match works when your conversion tracking is accurate, your primary conversion action is correct, and someone reviews the search terms report weekly. Without those three, it spends fast on irrelevant traffic. I treat broad match as something an account earns once the signal is clean, not as a default setting on day one.

Is Performance Max worth running for a small business?

For ecommerce with a real product feed and steady conversion volume, often yes, with brand exclusions applied and a separate brand search campaign. For small lead-generation budgets, usually not. A thin conversion signal makes PMax chase the cheapest form fill available, and a well-negatived search campaign gives you far more visibility into what you are buying.

What is offline conversion import and why does it matter?

It sends the outcome of a lead back to Google Ads. The click identifier is captured on your form, stored in your CRM, then uploaded when the deal is qualified or won. The account then optimizes toward leads resembling your actual customers instead of toward form volume. For any business that closes by phone or sales call, it is the highest-leverage setup available.

How long before a restructured Google Ads account performs?

Expect a learning period after any structural or bid strategy change, during which results are unstable and should not be judged. I avoid changing bid strategies frequently for that reason. What you get quickly is diagnosis: within the first week you should know exactly what is broken and roughly what it has been costing you.

Do you work with businesses outside Nepal?

Yes. My clients are primarily overseas, and I delivered SEO and paid ads remotely for Minnesota-based US clients on contract with SoloiTech from December 2025 to June 2026. Work happens during your overnight hours, with a fixed overlap window for calls, and every account change is documented in writing rather than relayed verbally.

Google Ads & PPC