SEM Keywords: Stop Bidding on What You Already Own

SEM keywords are the search terms you bid on in paid search campaigns, but the real decision isn’t which words to buy. It’s understanding which searches represent genuine commercial opportunity versus which ones you’d win anyway, with or without the spend. That distinction shapes everything from budget allocation to campaign architecture to how you measure success.

Most SEM accounts I’ve audited over the years have the same structural problem: they’re built around capturing demand that already exists, from audiences already primed to convert, with budget weighted toward the terms most likely to show a clean return in the dashboard. It looks efficient. It often isn’t growth.

Why Most SEM Keyword Strategies Are Built Backwards

When I ran iProspect UK, we managed significant paid search budgets across dozens of clients. One pattern showed up repeatedly: accounts that had been optimised for efficiency over years often had a shrinking addressable audience. Click-through rates looked strong. Conversion rates looked strong. But the volume of new customers was flat or declining, because every efficiency gain had come from tightening the funnel rather than widening it.

The problem starts with how keyword strategies get built. Most start with a seed list, run it through a keyword planner, sort by volume, and then filter down by estimated CPC and competition. The result is a list that looks rational but is actually a map of existing demand, not a map of commercial opportunity. You end up bidding on the same terms as every competitor, fighting over the same searchers, in the same auction, at the same time.

That’s not a strategy. That’s a participation fee.

The better starting point is a question: which searches represent people who don’t already know us, haven’t already decided to buy from us, and are at a stage in their thinking where paid search can actually change their behaviour? That’s a much smaller list than the keyword planner produces, but it’s a far more commercially valuable one. If you’re thinking about how SEM fits within a broader go-to-market framework, the Go-To-Market and Growth Strategy hub covers the wider context in detail.

Branded Keywords: The Comfortable Budget Trap

Branded Keywords: The Comfortable Budget Trap

Branded keyword bidding is one of the most persistent debates in paid search, and I’ve sat on both sides of it. The argument for bidding on your own brand terms is usually framed around defending real estate, controlling the message, and blocking competitors. All of those are legitimate concerns in some contexts. But they’re often used to justify spend that flatters the account rather than drives the business.

Here’s the honest version: if someone searches your brand name, they already know you exist. They may already be a customer. They may already be on their way to your site. When you bid on that term and they click your paid ad instead of your organic listing, you’ve paid for a conversion that was almost certainly going to happen anyway. The attribution model credits the keyword. The business is no better off than it would have been without the spend.

This isn’t an argument against all branded bidding. There are genuine cases where it makes sense: when competitors are actively bidding on your brand terms, when you’re running a promotion that the organic listing doesn’t reflect, or when you’re testing messaging in a controlled way. But branded keywords should be a deliberate, justified line item, not a default that quietly absorbs 30-40% of the SEM budget because it keeps the reported conversion rate looking healthy.

I’ve seen this dynamic play out in client reviews more times than I can count. Strip out branded and retargeting from the performance report, and the picture changes substantially. That’s not a reason to panic. It’s a reason to be honest about what the SEM spend is actually doing.

Match Types and the Architecture of Intent

Match type management is where a lot of SEM budgets quietly bleed. Broad match, particularly since Google’s repeated expansions of what “broad” actually means, can serve your ads against searches that have a loose semantic relationship to your keywords but no meaningful commercial connection to your offer. Without rigorous negative keyword management, you end up paying for traffic that was never going to convert.

The standard advice is to use a tiered match type structure: exact match for your highest-value, most proven terms; phrase match for controlled expansion; broad match only where you have enough conversion data and budget headroom to let the algorithm learn without bleeding cash in the process. That’s broadly right, but it misses the more important point.

Match type decisions should be driven by what you’re trying to learn, not just by what you’re trying to control. Broad match, used deliberately with proper negative keyword scaffolding, is a legitimate discovery tool. It can surface search queries you hadn’t considered, reveal how your audience actually describes their problem, and identify adjacent categories worth targeting. The mistake is treating it as a set-and-forget setting rather than an active research function.

Negative keyword lists are where the real work happens. Every account I’ve inherited over the years has had gaps in negative keyword coverage that were costing real money. Not through any single catastrophic term, but through hundreds of small mismatches that individually look minor and collectively represent a significant efficiency drag. Auditing search term reports weekly, particularly in the first 90 days of a campaign, is not optional. It’s the job.

Competitor Keywords: The Underused Growth Lever

If branded keywords represent demand you already own, competitor keywords represent demand you’re trying to take from someone else. That’s a fundamentally different commercial proposition, and it requires a fundamentally different approach.

Bidding on competitor brand terms is legal in most markets (with some nuance around trademark rules and ad copy restrictions), and it can be effective, but the economics are often worse than people expect. Quality scores tend to be lower because your landing page isn’t optimised for the competitor’s brand name. CPCs tend to be higher for the same reason. And conversion rates are typically lower because someone searching for a competitor is, by definition, not yet searching for you.

The more interesting opportunity is competitor category terms: the non-branded searches that describe what your competitor does, the problems they solve, the products they sell. These searches represent people who are in-market, aware of the category, but haven’t yet committed to a specific brand. That’s where you have a genuine chance to change the outcome, rather than intercepting someone who’s already made a decision and hoping they’ll switch.

Earlier in my career I was more focused on lower-funnel performance metrics than I should have been. The numbers looked clean. Conversion rates were strong. But I was largely capturing intent that already existed, not creating it. The growth that actually moved the needle came from reaching people earlier in their thinking, before they’d formed a preference. Competitor category terms sit in that space. They’re harder to convert, but they’re reaching genuinely new audiences rather than recycling existing ones. This is a point that Semrush’s analysis of market penetration strategies touches on, though the SEM application is often underplayed.

Long-Tail Keywords and the Volume Trap

Long-tail keywords get a lot of attention in SEM strategy discussions, usually framed around lower CPCs, less competition, and higher purchase intent. All of that is directionally true. But the long-tail conversation often obscures a more important question: does the volume justify the complexity?

Managing hundreds of highly specific, low-volume keywords creates account complexity that has a real cost. It requires more time to manage, more budget to test, and more data to optimise. For large accounts with dedicated SEM teams and significant budgets, that complexity can be worth it. For smaller accounts, it often isn’t. The marginal return from the 200th long-tail keyword is almost always lower than the return from better creative, better landing pages, or better bid management on the 20 terms that already drive most of the volume.

The more useful framing for long-tail keywords isn’t volume versus competition. It’s specificity of intent. A search for “running shoes” tells you very little about what someone is actually trying to do. A search for “trail running shoes for wide feet under £100” tells you a great deal. The value of the long-tail isn’t that it’s cheaper. It’s that it’s more informative, both for targeting and for understanding what your audience actually needs.

When I was working with retail clients, the search term data from long-tail queries was often more useful as a product and messaging insight than as a direct revenue driver. What people search for in specific, unprompted language is a window into how they think about their problem. That’s valuable regardless of whether you’re bidding on the term.

Search Intent and the Funnel Mismatch Problem

Search intent is the framework that’s supposed to solve keyword strategy: informational, navigational, commercial, transactional. Map your keywords to intent categories, match them to funnel stages, and you have a coherent architecture. In practice, it’s messier than that.

The problem is that intent categories describe what someone is doing in a search session, not where they are in a buying process. Someone searching “best project management software” might be a researcher doing a market overview, a procurement manager building a shortlist, or a user who’s already using your product and looking for confirmation they made the right choice. The keyword looks the same. The commercial value is completely different.

This is where the connection between SEM keyword strategy and broader audience understanding becomes critical. If you know who your buyers are, how they make decisions, and what triggers their search behaviour, you can build keyword architecture that reflects actual buying patterns rather than a theoretical funnel. If you don’t have that understanding, you’re making assumptions that may or may not hold, and your keyword strategy will reflect those assumptions whether you’re aware of it or not.

I’ve judged the Effie Awards, and one thing that consistently separates effective campaigns from ones that just look efficient is the quality of the audience insight underpinning the strategy. The best SEM work I’ve seen isn’t characterised by clever bidding tactics. It’s characterised by a clear, specific understanding of who is searching, why, and what they need to see to take the next step. The keyword list is almost a by-product of that understanding.

Keyword Research Tools: Useful Data, Not Ground Truth

Every SEM practitioner uses keyword research tools. Google Keyword Planner, SEMrush, Ahrefs, and others all provide volume estimates, competition data, and related term suggestions. They’re useful. They’re also, without exception, an approximation of reality rather than a precise measurement of it.

Volume estimates vary significantly between tools. The same keyword can show 1,000 monthly searches in one tool and 5,000 in another. Neither is wrong exactly. They’re both modelling the same underlying data with different methodologies and different panel sizes. The practical implication is that keyword research tools should inform prioritisation, not determine it. Use them to understand relative volume and relative competition. Don’t treat the absolute numbers as facts.

The most reliable keyword data you have access to is your own search term reports. What people actually searched for when they clicked your ad is real data, not modelled data. It reflects your specific audience, in your specific market, at this specific point in time. Keyword research tools can tell you what searches exist. Your own account tells you what searches are commercially relevant to your business.

There’s a broader principle here that I come back to repeatedly: analytics tools are a perspective on reality, not reality itself. That applies to keyword volume data as much as it applies to attribution models or social listening platforms. The tool is a starting point for thinking, not a substitute for it.

Negative Keywords: The Strategy Most Accounts Underinvest In

If positive keywords are about choosing what to target, negative keywords are about choosing what to exclude. Most accounts treat negative keywords as a cleanup task rather than a strategic function. That’s a mistake.

A well-built negative keyword list does several things simultaneously. It improves relevance by preventing your ads from showing against searches that don’t match your commercial intent. It improves efficiency by concentrating budget on searches more likely to convert. And it improves data quality by removing noise from your performance reports, so you can see more clearly what’s actually working.

The categories of negative keywords worth thinking about systematically are: job-related searches (people looking for employment, not products), informational searches where purchase intent is absent, competitor terms you’ve explicitly decided not to target, and geographic or demographic modifiers that don’t match your audience. None of these are exotic. All of them show up in search term reports with surprising regularity, and all of them represent budget that could be working harder elsewhere.

One pattern I’ve seen in agency audits: accounts that have been running for several years with minimal negative keyword maintenance often have a long tail of irrelevant searches that individually look trivial but collectively represent 10-15% of spend. That’s not a small number at any meaningful budget level. It’s also entirely preventable with a disciplined review process.

Keyword Strategy Across the Buying Cycle

One of the more useful ways to think about SEM keyword architecture is to map it explicitly against the buying cycle, not the marketing funnel. The funnel is an internal model. The buying cycle reflects how customers actually behave.

Early in a buying cycle, searches tend to be broad and exploratory. Someone is aware of a problem or a need but hasn’t yet formed a view on solutions. These searches are harder to convert directly, but they’re reaching people before preferences have formed. That’s commercially valuable, even if the attribution model doesn’t credit it appropriately.

Mid-cycle searches are more comparative. “Best X for Y”, “X versus Y”, “X reviews” type queries. The person knows what category they’re buying in and is evaluating options. This is where strong positioning and clear differentiation matter most in the ad copy and landing page experience.

Late-cycle searches are the transactional queries that most SEM accounts over-index on. “Buy X online”, “X price”, “X discount code”. High intent, high competition, expensive. These are worth bidding on, but they should be the end of the keyword strategy, not the whole of it.

The analogy I use with clients is a clothes shop. Someone who walks in and tries something on is far more likely to buy than someone who walks past the window. But the window still matters. It’s what got them through the door. SEM keyword strategy that only targets the equivalent of “I’m ready to buy right now” is missing the earlier moments that shaped that readiness. BCG’s work on brand and go-to-market strategy makes a similar point about the relationship between brand investment and commercial outcomes, though the SEM application is rarely drawn out explicitly.

Integrating SEM Keywords with Broader Go-To-Market Strategy

SEM keyword strategy doesn’t exist in isolation. It’s one component of a go-to-market approach, and the keyword decisions you make should reflect the broader strategic choices about who you’re targeting, what you’re offering, and how you’re positioning against alternatives.

If your go-to-market strategy is focused on a specific segment, your keyword strategy should reflect that specificity. Generic category terms might drive volume, but if that volume doesn’t match your target segment, it’s not serving the strategy. Conversely, if your growth objective is to expand into a new segment or category, your keyword strategy needs to include terms that reach people in that new space, even if the initial conversion rates are lower than your existing campaigns.

The connection between SEM and positioning is also underappreciated. Your ad copy is, in effect, a positioning statement delivered at the moment of highest commercial intent. If your positioning is unclear or undifferentiated, that will show up in your click-through rates and conversion rates before it shows up anywhere else. SEM is often the fastest feedback loop you have on whether your positioning is working.

I’ve seen this play out in turnaround situations. When a business is struggling commercially, the SEM account is often one of the first places the symptoms appear: declining quality scores, rising CPCs, deteriorating conversion rates. Not because the SEM management is poor, but because the underlying commercial proposition isn’t strong enough to compete. Fixing the keyword strategy won’t solve that. But the keyword data can point you toward what needs fixing. For a more complete picture of how SEM fits within growth strategy, the Go-To-Market and Growth Strategy hub covers the strategic frameworks that sit above channel-level decisions.

What Good SEM Keyword Strategy Actually Looks Like

After 20 years of managing, auditing, and advising on SEM accounts across 30 industries, the accounts that perform consistently well share a small number of characteristics that have nothing to do with platform features or bidding algorithms.

They have a clear rationale for every keyword in the account. Not “it came up in the keyword planner” but a specific commercial reason why this search term represents an audience worth reaching at this point in their buying process. That sounds obvious. It’s rarer than it should be.

They treat the search term report as a primary data source, not a secondary audit task. What people actually search for is more informative than what you predicted they’d search for, and the gap between those two things is where most of the optimisation opportunity lives.

They allocate budget with a view on what they’re trying to achieve commercially, not just what’s performing well by platform metrics. A campaign that’s generating cheap conversions on branded terms might be less valuable than a campaign generating expensive conversions on competitor terms, if the latter is bringing in genuinely new customers. Semrush’s overview of growth examples illustrates how channel-level decisions connect to growth outcomes, though the SEM specifics vary by context.

And they connect keyword strategy to audience understanding. The best keyword lists I’ve seen aren’t built from tools. They’re built from knowing, specifically, who your buyers are, what language they use, what problems they’re trying to solve, and where in their thinking process they are when they search. The tools help you validate and expand that list. They don’t replace the thinking that produces it.

Paid search, at its best, is a direct line to commercial intent. The keyword strategy you build determines whether you’re standing in the right place when that intent shows up, or whether you’re bidding on the comfortable, familiar terms that make the dashboard look good while the business stays flat.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions