The keyword landscape has shifted beneath our feet
We are tracking a clear inflection point in how wiki:keyword-research works across both the Apple App Store and Google Play. The dynamics that once rewarded broad, trending keyword placement — think "free" in 2018, "AI" today — are giving way to a market that punishes vagueness and rewards precision. Three converging trends are driving this shift: keyword saturation at the category level, softening demand on previously reliable high-volume terms, and the growing importance of locale-specific research over translation-based approaches.
For ASO practitioners, the message is unmistakable: the era of lazy keyword strategy is over.
"AI" is the new "free" — and it is already losing its edge
In six of more than twenty major App Store categories, "AI" has become the single most-used keyword in app metadata. In Productivity, it now outranks "notes," "tasks," and "daily." In Photo & Video, it sits above "photo," "video," and "camera." In Entertainment — and this is the striking one — "AI" now outranks "TV."
Two distinct dynamics are at play:
- AI as a genuine utility signal. In categories like Productivity and Photo & Video, the keyword describes real, differentiated capabilities: on-device processing, generated outputs, adaptive personalization. When a design app leads with "AI Video & Photo Editor," it is describing something tangible that users can find immediately inside the product.
- AI as a credibility badge with no substance. In Entertainment, Lifestyle, and increasingly Health & Fitness, "AI" is being used as a quality signal without underlying specificity. It means "our algorithm is smart" or "we have better recommendations" — things that are invisible to users and indistinguishable across competitors.
Even specific AI compound keywords — "AI calorie counter," "AI photo editor," "AI habit tracker" — carry lower search volume than their non-AI equivalents while having comparable keyword difficulty. The saturation has made the modifier a net negative in many contexts.
Broad demand is softening — and that changes everything
The mental health app category offers a useful case study. Several core search terms that powered growth two years ago — terms like "therapy," "meditation," "anxiety" — have moved downward in daily impression volume on US iPhone. The underlying need is as real as ever (public health data makes that clear), but app store search demand is not tracking upward across the board.
Download trends across major apps in the category tell a similar story: five of six tracked apps with full data windows declined in estimated downloads over an eleven-month period. This is not a category in freefall, but it is no longer a rising tide that lifts average execution.
When demand was racing upward, mediocre keyword strategy could still look clever. When demand softens, mediocre strategy gets expensive fast. The teams winning now are those who:
- Track keyword demand proxies with proper nuance, not just monthly check-ins
- Show up in search where intent is commercially meaningful, not just high-volume
- Use Custom Product Pages to match intent after the tap — routing anxiety-related queries to pages that lead with reassurance, and meditation queries to pages that emphasize routine and content depth
The fundamentals still hold, but the bar is higher
The core process of wiki:keyword-strategy has not changed. What has changed is the sophistication required to execute it well.
Building a semantic core
The foundation remains the same: collect as many relevant search requests as possible, then ruthlessly prioritize. A well-built semantic core typically starts with 500+ raw keywords and narrows to roughly 200 per country after filtering out terms with negligible search volume (popularity scores below 15 on most tools).
The collection process should pull from multiple sources:
- Existing index positions. Identify keywords your app already ranks for, especially those in positions 5–20 where a small push could move you onto page one.
- Competitor keywords. Analyze what your keyword competitors (not necessarily your business competitors) rank for. A meditation app competes for keywords with sleep sound apps, breathing exercise tools, and yoga timers.
- Autocomplete and suggest tools. The alphabet technique — typing your seed keyword followed by each letter — remains one of the simplest and most reliable methods for uncovering long-tail opportunities.
- Search Ads recommendations. Platform-provided keyword suggestions for paid campaigns reflect high-volume, high-relevance queries that are equally valuable for organic targeting.
- User reviews and support tickets. Real users describe your app in their own language. If multiple reviews call your product a "habit tracker" and your metadata never mentions that phrase, you are leaving keywords on the table.
- Search volume: How many users search for this term monthly. A score of 40–50+ on standard ASO tool scales generally indicates meaningful traffic potential.
- Keyword difficulty: How competitive the top results are. The sweet spot is moderate-to-high volume with low-to-moderate difficulty — achievable opportunities with real traffic.
- Relevance: The most overlooked metric. A keyword with massive volume and low difficulty is worthless if users who find your app through it immediately uninstall. High bounce rates send negative signals to the algorithm that hurt rankings across the board.
Platform-specific mapping
Apple and Google take fundamentally different approaches to indexing, and keyword strategy must account for these differences:
- Keyword Field (100 chars): hidden backend field, comma-separated
- Description: not indexed for search
- Golden rule: never duplicate keywords across title, subtitle, and keyword field — Apple treats all three as a combined set
- Full Description (4,000 chars): fully indexed via NLP — keyword placement and density matter
- Additional signals: backlinks, review content, engagement metrics, developer name
Localization is keyword research, not keyword translation
This is the single most damaging error we see in international ASO: teams translate their English keywords instead of researching local ones from scratch.
The numbers make the case. The English keyword "workout tracker" may generate 45,000 monthly searches in the US App Store. Its direct German translation, "Trainingstracker," pulls only 800 searches. The actual high-volume term German users search for — "Fitness Tagebuch" (fitness diary) — gets 12,000. By translating instead of researching, you optimize for a keyword with 15x fewer searches. Multiply that across 10–20 markets, and the cumulative download loss is staggering.
The same pattern repeats across languages. Japanese users search for the equivalent of "mind calming app" rather than a literal translation of "meditation app." German users type "Bildbearbeitung" (image processing) rather than a translated "photo editor."
Additional localization pitfalls that compound keyword errors:
- Character limit overflows. German compounds and Finnish words routinely exceed English character counts. Truncating mid-word looks unprofessional; rewriting to fit requires actual local-language fluency.
- Untranslated screenshots. Localized metadata with English-only screenshots creates an immediate trust disconnect. Localized screenshots lift conversion rates by 20–30% in non-English markets.
- Ignoring RTL languages. Arabic, Hebrew, Persian, and Urdu require full right-to-left layout support — not just mirrored text direction but adapted navigation, alignment, and icon placement.
What to do right now
For teams reassessing their keyword strategy in light of these shifts, here is a prioritized action list:
- Audit your "AI" usage. If you are using the keyword as a credibility badge rather than a descriptor of a tangible, differentiated feature, test removing it. Measure conversion impact over a full update cycle.
- Re-evaluate high-volume terms for demand trends. Keywords that powered growth two years ago may be softening. Track impression-level data, not just rankings.
- Segment intent and route accordingly. Use Custom Product Pages (Apple) and Custom Store Listings (Google) to match post-tap messaging to the specific intent behind each keyword cluster.
- Stop translating keywords. Invest in local-language keyword research for every market you serve. The ROI difference between translated and researched metadata is often an order of magnitude.
- Expand your long-tail coverage. In saturated categories, specific multi-word queries with clear intent outperform generic high-volume terms on both conversion and ranking achievability.
- Never duplicate across Apple metadata fields. This remains one of the most common wastes of character space. Title, subtitle, and keyword field are a combined set — every repeated word is a missed opportunity.