Remember when “exact match” meant… exact? Pepperidge Farm remembers. Google, meanwhile, has been busy quietly swapping your tidy keyword/match-type universe for a fuzzier world of “themes,” “intent,” and “let the machines figure it out.” Translation: less manual control, more machine learning deciding who sees your ads, when, and why.

If that gives you a tiny marketer panic attack, breathe. This shift can work for you — if you stop fighting yesterday’s rules and start playing today’s game.

What changed (and why your old playbooks keep breaking)

  • Exact and phrase match have been “relaxed”: Close variants, paraphrases, and “same intent” matches sneak in like that mate who always shows up to parties with four extra people.
  • Broad match got a protein shake: With smart bidding, broad match isn’t the budget-vaporizer it used to be. It’s now Google’s favorite data pipeline for feeding its models more context.
  • RSAs and auto-expanding themes: Fewer levers to micro-control. More signals for Google to test permutations, audiences, and placements you didn’t hand-pick.
  • The big picture: Google wants to optimize toward outcomes (conversion value) rather than your handcrafted list of words. It’s intent > syntax.

Net effect: Your account structure matters less than your signal quality.

Stop obsessing over words. Start engineering signals.

Old world: build 58 ad groups for 58 micro-variants. New world: give the algorithm a clear objective, clean conversion tracking, and enough diversity in creative to test. Then referee it with guardrails.

  • Track the right thing (accurately)
    • Ditch soft goals. Optimize to real revenue or qualified leads, not page views or “spent 30s on site.”
    • Use Enhanced Conversions, server-side or GTM server where possible. Dirty data = drunk bidding.
  • Feed it meaningful conversion value
    • Lead gen: apply weighted values (SQL > MQL). Even better: offline conversion imports with actual revenue.
    • Ecommerce: use dynamic values and include returns data if you can.
  • Consolidate where it helps
    • Fewer, stronger campaigns with enough volume beat a sprawl of starving ad groups.
    • Group by intent/theme, not pedantic keyword variants.

Broad, but not blind: how to use intent without torching budget

  • Pair broad match with smart bidding and guardrails
    • Start with tCPA/tROAS targets informed by reality, not hope.
    • Layer audiences: remarketing lists, custom segments (searchers of X, site visitors of Y).
  • Negative keywords are your new superpower
    • Weekly search term audits. Add negatives aggressively (brand safety, competitor brands you won’t touch, unqualified “jobs,” “free,” “definition,” etc.).
    • Build negative libraries by theme and share across campaigns.
  • Theme-first structuring
    • Example (B2B SaaS): “Pain theme” (problem queries), “Solution theme” (category queries), “Brand” (protect), “Competitor” (surgical). Each with its own creative angle and bid target.

Creative matters more now (finally)

If Google’s expanding reach via “intent,” your ads must do the filtering. Write for the person, not the word.

  • Message-match the intent, not the keyword
    • Pain theme: lead with outcomes and empathy.
    • Solution theme: clarity and proof (ratings, ROI, case studies).
    • Competitor theme: a sharp, respectful contrast.
  • Use RSAs with intent-specific assets
    • 8–12 headlines mixing pain points, value props, proof, qualifiers.
    • Pin sparingly; let testing happen, but keep must-have claims pinned 1–2.
  • Landers that pre-qualify
    • Specific pricing cues, ICP callouts, social proof, and clear next steps.
    • Speed: sub-2s or pay the tax.

The four levers that still move mountains

  1. Target and signals
  • Clean conversion setup, correct attribution window, values in place.
  • Smart bidding with realistic floors and ceilings.
  1. Query control
  • Negatives, brand safety, geo trims, hour/daypart rules based on actual conversion hours.
  1. Creative and offer
  • People convert on offers, not algorithms. Test real differences (trial length, bonus, guarantee), not button color.
  1. Measurement and feedback loops
  • Import offline conversions; pass GCLID/GBRAID back to ad-level. Shorten the learning feedback cycle.

How to migrate without faceplanting

  • Phase, don’t flip
    • Keep your best-performing exact/phrase “legacy lanes” live.
    • Stand up parallel intent-themed campaigns with broad+smart bidding.
    • Cap budgets at 10–20% to learn, then reallocate as performance proves out.
  • Define your red lines
    • CPA/ROAS guardrails, impression share on brand, and a weekly search term SLA.
  • Run incrementality checks
    • Geo splits or time-based tests to ensure you’re growing net-new, not just shuffling last-click credit.

When the machine goes rogue

Even the best models wander. Here’s your “tap the glass” checklist:

  • Sudden CPC lift, flat CVR? Tighten negatives, raise value signals, trim geos/hours.
  • RSAs stagnating? Swap in 3–4 fresh headlines and 2–3 descriptions tied to a new proof angle.
  • Learning mode purgatory? Increase daily budget slightly, consolidate skus/ad groups, or broaden audience signals.

The uncomfortable but freeing truth

Keywords were training wheels. Useful, familiar, sometimes comforting. But intent-first + ML is where Google Ads lives now. If you cling to match-type micromanagement, you’ll keep losing to brands that engineer better signals, offers, and creative — even if their keyboard is covered in coffee.

So yes: let the machine “expand.” Your job is to:

  • Define winning (with accurate, value-based conversions).
  • Set lanes (negatives, budgets, geos, brand protection).
  • Show up with a brand and offer people actually want.

Do that, and intent-driven automation stops being a black box and starts being your unfair advantage.

Chris Barnard has spent over 15 years delivering exceptional revenue growth for ambitious businesses in the UK, Europe and North America through his marketing technology business, FeedbackFans.com and as an independent business consultant.

By his mid-20’s he was running digital departments for FTSE100 companies in London, eventually leading to a very successful period in digital customer acquisiton for a well-known brand in his early 30’s generating nine-figure revenues with seven-figure budgets. He now puts his experience, knowledge and ideas into good use, supporting challenger insurgent brands and forward thinking businesses to outperform in their sectors, whilst disrupting and improving the marketing, technology and development sectors.

Feedback Fans provides a unique next-generation managed technology and marketing platform that delivers outstanding and out-sized results for businesses in sectors such as finance, retail, leisure, and professional services.

With our unparalleled expertise in creating cutting-edge solutions and environments, we empower our clients and users to thrive and outperform in the digital age.

Chris Barnard is Managing Director of FeedbackFans.com and producer of the Bear Business Vodcast