How Not to Burn Budget on Google Ads

Broad keyword match is a powerful tool, but when you use it without managing a negative-keyword list, the campaign starts responding to irrelevant queries.

I once saw a renovation contractor's campaign get clicks on "photo retouching" and "car detailing." Every click cost money, and not one was a potential customer.

Some campaigns work great in the morning and waste money at night. If all your conversions come in between eight and six, there's no point paying for clicks at two in the morning.

You write one ad, set it up, and wait. That's the common pattern. But in most cases, you can't know in advance which headlines will work.

A headline that emphasizes price might beat one that emphasizes professionalism, or the other way around — it depends on the field and the audience. Google offers RSAs that assemble different combinations, but it's still worth testing variations.

This cuts both ways. Some people touch the campaign every day and never give the algorithm time to learn. Others set it up and forget it.

Budget depends on the competition in your field, but you need at least 5–10 clicks a day to have data for learning. In a field with an average CPC of ₪15, that means at least ₪2,250–₪4,500 a month — just to gather basic data, before any optimization.

In competitive fields like law, real estate, or medical aesthetics, a budget of ₪5,000 and up is the realistic minimum. Entering at ₪500 a month in fields like these won't generate enough volume for the algorithm to learn, and it mostly just wastes money.

A new campaign needs 2–4 weeks for initial learning. During this phase the algorithm tests which audiences, hours, and devices perform — and you mustn't make big changes that reset it to square one.

Data you can rely on only arrives after 60–90 days. Anyone who shuts a campaign down after three weeks "because it isn't working" usually shut it down right before it started to work.

On low budgets (under ₪5,000) you can manage it yourself — mistakes are less costly and there's value in the hands-on learning. But above that, professional mistakes (Broad Match with no negatives, wrong bidding, broken measurement) can cost more than you saved on a management fee.

The common problem with self-management isn't knowledge — it's time. A campaign left unchecked for two weeks quietly deteriorates. With a budget over ₪15,000 a month, a professional campaign manager pays for themselves many times over.

Google relies on search intent — the searcher is already looking for something specific, and you show up at exactly that moment. Meta relies on a user profile and demand creation — you reach people before they've even thought they need you.

For services with existing demand (people are searching on Google), Google wins. For new products people aren't searching for on their own, Meta is a better fit. The strongest combination: Google for existing demand + Meta for audience building and remarketing.

Usually not. A small business needs transparency and control — to know exactly what's being searched, what's clicking, and what's converting. PMax runs ads across all of Google's networks but provides minimal reporting, which makes it very hard to know where the money is going.

PMax also requires a large volume of data to work properly — at least 50–100 conversions a month. The recommendation: start with a Search campaign, build up conversion data, and then consider PMax once you have a foundation. Not the other way around.

AI Overviews — direct AI answers at the top of the search page — lower CTR on informational searches (what is, how to). The searcher gets an answer without clicking at all. This forces a focus on searches with higher purchase intent, where ads still deliver clear value.

Google's own bidding and creative tools are also becoming AI-driven — which means you have to feed them better conversion data. A 2026 campaign without proper measurement is a campaign whose AI is running blind.

Yes. The Google Ads Transparency Center shows the active ads of any advertiser — you can search by business name and see what they're running right now. It's a free tool, accessible to everyone.

External tools like Semrush, SpyFu, and SimilarWeb allow a deeper analysis: which keywords your competitors are bidding on, the estimated cost per click, and which ad copy they're choosing. That's intelligence worth its weight in gold before launching a new campaign.

The price varies dramatically by field: lawyers and medicine — ₪15–₪60 per click. Insurance and finance — ₪8–₪25. Renovation and construction — ₪5–₪15. Less competitive local fields — ₪1–₪6.

What affects the price in practice: your Quality Score (a high score = a lower price), the level of competition on the specific phrase, and landing page relevance. Google doesn't just calculate who bids the most — it also looks at who delivers the best experience to the searcher.

The common reasons are: a mismatch between the ad's promise and the landing page content (the user expected one thing and got another), an unclear value proposition on the page, a form that's too long and complicated, and technical issues — slow mobile loading that causes people to abandon before the page even opens.

Sometimes the problem is the keywords themselves — if you're attracting an audience looking for information rather than a service, the clicks will come but the conversions won't. You need to examine the entire visitor chain: what they searched, what they saw in the ad, which page they reached, and what action you asked them to take.

In e-commerce campaigns, a ROAS of 3–4 (a return of ₪3–4 for every shekel invested) is a reasonable starting point — but it depends entirely on the product's gross profit margin. A product with a 20% margin needs a far higher ROAS than one with a 60% margin.

In lead campaigns (not direct sales), ROAS isn't the right metric — you need to calculate cost per lead (CPL) and from there customer acquisition cost (CAC) relative to a customer's lifetime value (LTV). Without those numbers, any "ROAS target" is a guess.

Only after the campaign has accumulated at least 30–50 conversions in the past month. Without enough baseline data, the algorithm runs on guesswork — and can waste a very large budget without converting while it "learns."

The switch should be gradual: first Maximize Conversions (no CPA target), and after the algorithm has accumulated around 30 conversions, add a CPA target based on the data already gathered. A CPA target set before there's any data is a recipe for failure.

Quality Score (1–10) is set by three factors: the ad's relevance to the keyword, the landing page experience, and the expected click-through rate (CTR) compared to other advertisers on the same word. A high score = a lower cost per click and a better position.

What to do: make sure the keyword appears in the ad headline, lead to a relevant, fast landing page that contains the phrase, and check that the historical CTR on the specific keyword isn't unusually low. Every additional point in the score can save 15–20% on the cost per click.

Broad Match lets Google show ads on "related" searches — at Google's own discretion. In practice, that means an ad for the phrase "divorce lawyer" could also show up for a search like "books about divorce" or "free family counseling."

Without tight management of negative keywords and a weekly search terms report review, Broad Match can eat half your budget on an audience that doesn't convert. If you use it, do so only in combination with Smart Bidding and an up-to-date negative-keyword list.

The homepage is meant for people who already know you and want to navigate. A visitor who searched "emergency plumber in Tel Aviv" and lands on a renovation company's homepage loses context, gets frustrated, and leaves within seconds. This hurts CTR, Quality Score, and conversion rate alike.

Every service or product in the campaign needs a dedicated landing page that contains: the ad's promise, social proof (reviews), a clear offer, and a contact form above the fold. A focused landing page converts 2–5 times better than a generic homepage.

Four conversions are a must for any service business: form submissions, clicks on the phone number, WhatsApp clicks, and inbound phone calls (via Google Forwarding). Without them, the algorithm doesn't know what to bring more of and what to rule out.

Broken measurement is the most common mistake — a campaign that runs for months without proper measurement never gets a chance to learn. You need to verify that measurement works before you turn the campaign on, not after you've already spent the money.

Search intent is the reason behind the search: "what is life insurance" (informational) versus "life insurance price for age 40" (transactional). The chance of conversion on a transactional search is several times higher — but the CPC there is correspondingly higher too.

The common mistake: chasing high volume and bidding on informational phrases that don't convert. It's better to concentrate budget on 10 phrases with high purchase intent than on 50 phrases that attract visitors who just want to learn.

Three assets you can't skip: Sitelinks (links to additional pages), Callouts (short phrases that highlight benefits), and the call extension (Call Asset). Together they increase your ad's footprint on the results page and improve CTR by tens of percent.

Additional assets worth adding: Image Assets, Price Assets if relevant, and Structured Snippets. Google rewards advertisers who give it more "material" to work with — with a higher Quality Score and a lower cost per click.

A full website isn't required — but a dedicated, professional landing page is essential for the campaign to succeed. A page that includes: a clear headline matching the ad's promise, a service description, customer reviews, and a contact form — that's the minimum that will work.

A full website is worth having over time because it improves Quality Score and builds credibility, but you don't need it to launch a first campaign. Tools like Unbounce, Instapage, or Elementor let you build a professional landing page in just a few hours.

Three mistakes that recur in most campaigns: not reviewing the search terms report every week — Google spends money on searches you never approved. No negative keywords — without an up-to-date list, every Broad Match keyword is a hole in the bucket. Broken conversion tracking — a campaign that doesn't know what converted is a campaign that wastes.

Two more mistakes worth avoiding: changing your bidding strategy every week (every change resets the algorithm to the learning phase), and spreading budget across too many simultaneous campaigns without a minimum volume for each.

The true metric is customer acquisition cost (CAC) and bottom-line profitability — not cost per click and not the number of visits. A campaign with a ₪40 CPC that brings in a customer worth ₪5,000 is far better than a campaign with a ₪3 CPC that doesn't convert.

Questions worth asking every month: How many leads came from Google Ads? How many of them became customers? What's the average acquisition cost? Is it lower than the value a customer brings to the business? If the answer to the last one is yes, the campaign is working. If not, you need to diagnose where the chain breaks.