Do startup marketing in this order: measurement first, then SEO and answer engine optimization (AEO) on pages that answer real buyer questions, then app store optimization (ASO) if you have an app, then paid ads once you know your conversion rate. SEO still works in 2026: Google says its AI features rest on its core ranking systems, and Alphabet reports more than 2.5 billion monthly users of AI Overviews.
Order matters because the channels compound differently. Measurement makes every later decision cheaper. Search and answer-engine content keeps working for years. ASO keeps working for as long as the app is listed. Paid traffic stops the day the budget does. This guide explains each channel as it works in October 2026, then gives an order of operations in weeks and a map of what to measure.
What AEO and GEO mean
Answer engine optimization (AEO) is the practice of making content easy for AI answer engines, such as Google AI Overviews and AI Mode, ChatGPT and Perplexity, to find, trust and cite. Generative engine optimization (GEO) is the same idea under the name used in a research paper first published in November 2023 by researchers at Princeton and other institutions and presented at KDD 2024, which found that content changes could raise visibility in generative answers by up to 40%.
The audience is large. Alphabet said in June 2026 that AI Overviews have more than 2.5 billion monthly users and AI Mode more than 1 billion, and OpenAI said at DevDay on 29 September 2026 that ChatGPT reaches 1.2 billion people a week. Clicks are scarcer. Pew Research Center found that Google users who saw an AI summary clicked a traditional result on 8% of visits, against 15% when no summary appeared, and clicked a link inside the summary on 1% (March 2025 data). Being the cited source now matters alongside ranking.
AEO tactics that work
Google published guidance in May 2026 stating that its generative AI features rest on the same core ranking and quality systems as Search, and that SEO best practices remain the foundation. In practice that means crawlable pages, clear technical structure and original content drawn from first-hand experience. On top of that base, these tactics make a page easier to cite.
- Quotable passages. Put a 40 to 70 word direct answer under each question-shaped heading, written to stand alone if it is lifted out of the page.
- Verified, dated facts. Name the source and the month for every number, and update the page when the number changes. Specific, attributed claims are what answer engines quote.
- Tables for comparisons. Prices, limits and options in rows are easier to extract than the same facts in prose.
- Crawler access. Allow OAI-SearchBot, which OpenAI uses to surface sites in ChatGPT search, and PerplexityBot in robots.txt. GPTBot, OpenAI's training crawler, is a separate decision.
- Consistent company facts. Use the same name, location, founding year and services on your site, LinkedIn and directory profiles, with Organization schema on the site. Google's guidance says seeking inauthentic mentions across the web is not as helpful.
- Visible authorship and dates. Show who wrote the page, what experience they bring, and when the facts were last checked, as every post on this blog does.
FAQ schema: still useful, but not for rich results
FAQPage markup still helps machines parse question and answer pairs, and every post on this blog carries it. But Google stopped showing FAQ rich results in Search on 7 May 2026, and its AI guidance says structured data is not required to appear in generative AI features. Add it for clarity, not for a search feature that no longer exists.
llms.txt: cheap to add, ignored by Google
llms.txt is a Markdown file at the root of a site that lists its most useful pages for language models. Jeremy Howard of Answer.AI proposed it in September 2024. In June 2026 Google added a note that these files are not needed for Google Search and have no effect on visibility or rankings, though keeping one for other services is fine. Treat it as a ten-minute courtesy to the tools that read it, not as a ranking tactic.
How to check whether you are cited
Pick the 20 questions your buyers ask most. Once a month, ask each one in Google AI Mode, ChatGPT and Perplexity, and record which sources are cited and whether you are among them. Compare that log with impressions in Search Console's generative AI performance report and with AI referrals in GA4. When a competitor is cited and you are not, read their page: it usually answers the question more directly, with a more recent number.
ASO for an app launch
App store optimization (ASO) is search optimization for the App Store and Google Play: choosing the words your listing ranks for, and the screenshots that turn a view into an install. The constraints are tight and worth memorising.
- App Store: app name 30 characters, subtitle 30, keyword field 100. Spend the keyword field on terms not already in the name or subtitle.
- Google Play: title 30 characters, short description 80, full description 4,000. There is no keyword field, and Google's listing guidelines warn against lists of keywords in the description.
- Custom product pages: since 29 October 2025 Apple allows up to 70 per app, and a page with assigned keywords can appear in search results in place of the default page.
- Product page optimization: test up to three alternative App Store pages against the original before committing to new screenshots.
- Localisation: every language is a new keyword field and a new set of screenshots. SlimAI, the AI calorie app we built, ships in seven languages, and each one is a listing to optimise.
When paid makes sense
Paid ads are worth running once three things are true: a key event fires reliably in analytics, a landing page or store page already converts organic visitors, and you have a rough idea of how long a customer takes to pay back what it cost to acquire them. Before that, paid spend mostly buys data you cannot interpret.
- Start with intent. Search ads on problem-shaped keywords, and Apple Ads (renamed from Apple Search Ads in April 2025) on category terms, reach people who are already looking.
- Respect learning thresholds. Meta's guidance is that an ad set needs about 50 optimization events in a week to exit its learning phase; below that, optimise for an earlier event or consolidate ad sets.
- Cap and date every test: one channel, one offer, a fixed budget and a stop date, all decided before launch.
Order of operations, in weeks
| Weeks | Focus | What ships | Move on when |
|---|---|---|---|
| 0–2 | Measurement | GA4 key events, Search Console, App Store Connect and Play Console access, UTM rules, an AI-referral channel group | Every signup or purchase is recorded once, with its source |
| 2–6 | Foundations | Positioning, 5–10 core pages, technical SEO fixes, Organization schema, consistent profiles | Core pages indexed with no crawl errors |
| 4–12 | SEO and AEO content | 8–12 guides answering buyer questions, each with a direct answer, a table and an FAQ | Impressions rising in Search Console; first citations in AI answers |
| 6–10 | ASO, if you have an app | Keywords, screenshots, localised listings, first custom product pages | Store conversion rate measured by source |
| 8–16 | Paid tests | One channel, one offer, a capped budget and a stop date | Cost per key event known at a volume you trust |
| 16–26 | Scale what works | Refresh winning pages, add languages, add a second paid channel, PR | Payback period fits inside your runway |
What to measure, and where
| What to measure | Tool | Notes |
|---|---|---|
| Signups, trials and purchases | GA4 key events | Google renamed GA4 conversions to key events in March 2024; mark only the events that matter |
| Organic clicks, impressions and queries | Search Console Performance report | Keeps 16 months of data; set up the BigQuery export for longer history |
| Visibility in AI Overviews and AI Mode | Search Console generative AI performance report | Impressions only, no clicks; rolled out to all sites by 31 August 2026 |
| Referrals from AI assistants | GA4 traffic acquisition | ChatGPT adds utm_source=chatgpt.com to the links it cites; group AI referrers into one channel |
| App Store listing performance | App Store Connect Analytics | Impressions, product page views and conversion rate by source |
| Google Play listing performance | Play Console store listing reports | Store listing visitors, acquisitions and conversion rate by source |
| Paid efficiency | Ad platform plus GA4 | Cost per key event and payback by channel, checked against GA4 rather than platform reports alone |
How much to spend
Established companies spend about 7.8% of revenue on marketing in 2026, up from 7.7% in 2025, according to Gartner's CMO Spend Survey of 401 marketing leaders, which also found that CMOs put 15.3% of marketing budgets into AI initiatives. A pre-revenue startup has no revenue to take a percentage of, so budget in experiments instead: fund two or three paid tests with fixed caps and stop dates, keep measurement, content and ASO as the steady base, and raise spend only on channels whose payback fits inside your runway.
If you want this sequence run for your product, our digital growth team builds the measurement layer first, writes content designed to be cited, and will tell you which channels to skip.
Frequently asked questions
- What is AEO in marketing?
- AEO, or answer engine optimization, is the practice of making content easy for AI answer engines such as Google AI Overviews, ChatGPT and Perplexity to find, trust and cite. In practice it means direct 40 to 70 word answers under question headings, dated and sourced facts, comparison tables, crawler access for AI search bots, and consistent company facts across the web.
- How much should a startup spend on marketing?
- Established companies spend about 7.8% of revenue on marketing in 2026, according to Gartner's CMO Spend Survey. Pre-revenue startups should budget in experiments instead: two or three capped paid tests with stop dates, plus steady investment in measurement, content and ASO. Increase spend only on channels whose cost per customer pays back inside your remaining runway.
- Does SEO still work in 2026?
- Yes. Google's May 2026 guidance says its generative AI features rest on its core Search ranking and quality systems, so crawlable, useful pages still win. What changed is clicks: Pew found people clicked a result on 8% of visits when an AI summary appeared, against 15% without one. Optimise to be cited as well as clicked.
- What is ASO?
- ASO, or app store optimization, is the work of ranking an app in App Store and Google Play search and converting listing views into installs. It covers the app name, subtitle and 100-character keyword field on iOS, the title and descriptions on Google Play, screenshots, localised listings, custom product pages, and A/B tests of store pages.
About the author
Zoraiz Ejaz
Co-founder, Techparser
Zoraiz Ejaz is a co-founder of Techparser and leads its engineering and product practice. He has spent close to a decade designing, building and scaling mobile, web and AI products for startups and enterprise teams across health, fintech, payments, social and education, from first architecture and release pipelines through to launch and years of production support. He writes about how to scope, cost and ship software that lasts.
