Demand Analysis:
Keyword Research and Prompt Research
The demand analysis by digitalagenten combines keyword research for SEO and Google Ads with prompt research for AI answers. It also maps search volume, search intent, cost per click and AI questions along the customer journey.
As a result, you get a prioritized backlog with target pages and revenue value.
Demand analysis for SEO, SEA and GEO: one data foundation for Google, Ads and AI answers
- Keyword data provides search volume, trend and cost per click per topic, for example.
- Prompts contain questions, comparisons and buying criteria, because AI dialogs are longer than search queries.
- The customer journey therefore organizes both, so keywords and prompts can be compared per stage.
| Method | Guiding question | Data source | Result |
|---|---|---|---|
| Keyword research | What do users search for in Google, and how often? | SISTRIX, Google Search Console, Google Ads Keyword Planner | Keywords with search volume, search intent and CPC |
| Prompt research | What do users ask ChatGPT, Gemini and AI Overviews before they decide? | SISTRIX AI, Rankscale and questions from sales | Prompt set with journey stage and persona |
| Demand analysis | Where is the effort worth it because demand is highest? | all sources combined | Demand map with target page and revenue value |
Demand analysis as the foundation for benchmark, share of voice and missed opportunity analysis
The sum of demand and topics is the basis for the benchmark, and therefore for the missed opportunity analysis. In the online marketing consulting of digitalagenten, demand analysis is the third step, after the audit and the benchmark.
Only when the search volume of a market is known, for example for car insurance, do the impressions in Google Search Console turn into a market share.
Without this reference value, share of voice cannot be calculated.
Share of voice = impressions in Search Console ÷ search volume of the topic cluster
For example, the 14 German keywords on car insurance in the matrix below add up to 165,500 searches per month (SISTRIX, October 2026).
With an assumed 24,825 impressions in Search Console, share of voice is 15%.
This makes keyword research the basis of all content planning. Three values shape every content strategy:
- Search volume: how big demand is per topic
- Search intent: which page type serves the search
- Keyword difficulty and competition: how realistic a ranking is
Customer journey as the framework: compare keywords and prompts per stage
Search volume and prompt data are imprecise. Within the same stage of the customer journey, they can still be compared reliably, because keyword and prompt follow the same intent there.
- Keywords with known search volume first indicate how big a stage is.
- Prompts without known volume then indicate which questions go to AI in the same stage.
- Clusters per stage therefore give better estimates than single keywords.
Customer journey matrix: car insurance example from Germany
| Stage | Research | Orientation | Comparison | Selection | Purchase | Use & service |
|---|---|---|---|---|---|---|
| Job to be done | Before I register my car, I want to know which cover I need. | I want to understand the options, because my premium should go down. | Then I want the right plan at a fair price. | Also, the insurer must pay reliably after a claim. | After that, I want to buy fast and register right away. | Afterwards, I want to settle claims quickly. |
| Typical questions | Which car insurance do I need? Partial or comprehensive cover? | What does the no-claims class mean? Is an approved-repairer plan worth it? | Which car insurance is the cheapest? Which one covers the most? | Until when can I switch? How good is claims handling? | How do I get an eVB number? Can I buy online? | What to do after an accident? How do I report a claim? |
| Keywords (German) | teilkasko oder vollkasko (1,300)partial or comprehensive coverkfz versicherung für fahranfänger (550)car insurance for new driverswelche kfz versicherung (100)which car insurance | schadenfreiheitsklasse (14,900)no-claims classwerkstattbindung (550)approved-repairer clause | kfz versicherung vergleich (78,300)car insurance comparisonkfz versicherung berechnen (13,600)car insurance quotegünstigste kfz versicherung (2,900)cheapest car insurance | kfz versicherung wechseln (10,700)switch car insurancekfz versicherung kündigen (4,600)cancel car insurance | evb nummer (37,200)eVB number for the registrationkfz versicherung abschließen (350)buy car insurance | unfall was tun (250)what to do after an accidentkfz schaden melden (200)report a car insurance claim |
Prompts per stage: AI questions without and with a brand
| Stage | Research | Orientation | Comparison | Selection | Purchase | Use & service |
|---|---|---|---|---|---|---|
| Prompts without brand | “Which insurance do I need now that I have bought my first car, and roughly what does it cost?” | “Partial or comprehensive cover: which pays off if my small car is five years old?” | “Which car insurers are cheap for new drivers but still good when a claim happens?” | “Which insurers settle claims fastest, and what should I check before I switch?” | “How do I get an eVB number today if I want to register the car tomorrow?” | “How do I report a parking dent, and will I lose my no-claims class?” |
| Prompts with brand | “Is [Insurer A] worth it for new drivers if I only drive a little?” | “Does [Insurer A] offer a plan with an approved-repairer clause, and how much do I save?” | “[Insurer A] or [Insurer B]: who is cheaper if I drive a compact car?” | “How satisfied are customers with [Insurer A] when it comes to claims handling?” | “Can I buy online at [Insurer A] and get the eVB number right away?” | “What do I have to do at [Insurer A] if the windshield has a stone chip?” |
Example: car insurance in Germany. German keywords with monthly search volume: SISTRIX, October 2026. Prompts translated from German. Brands as placeholders.
AI dialogs during consideration, Google searches through purchase and service
The stages also apply to automotive, with test drive, purchase and workshop, and to healthcare, with finding a doctor, treatment and aftercare.
One pattern appears early: for “Kfz-Versicherung” (car insurance), SISTRIX records 2,325 topics in German AI dialogs, 92% of them in consideration. In Google, much of the demand falls on purchase and switching instead, for example “evb nummer” with 37,200 searches per month (SISTRIX, October 2026).
For the evaluation, we also map the stages to the stages of the SISTRIX prompt data. Research thus corresponds to awareness, orientation and comparison to consideration, selection and purchase to decision, use and service to retention.
Keyword research for SEO: search volume, search intent and ranking chance per topic
Seed keywords, long tail and questions from Search Console, sales and competitors
First, keyword research starts with seed keywords from services, sales and Google Search Console. SISTRIX and Google Trends then extend the list with synonyms, long-tail keywords and user questions.
Search intent know, do, visit and website per keyword
Every keyword gets a search intent, because it determines the right page type. Guides serve know searches, for example, while service pages cover do searches and location pages cover visit searches.
Keyword difficulty and competition in the SERP
The SERP also identifies which websites rank today and which page types Google Search prefers. As a result, this gives the keyword difficulty, meaning the realistic ranking chance of your website.
We also bundle terms with the same search results into one cluster, so your own pages do not compete with each other.
Keyword list and keyword strategy: from keyword ideas to content with traffic
- Keyword ideas: Keyword tools like SISTRIX deliver hundreds of ideas, therefore we filter them by relevance for your business and your market.
- Keyword list: Relevant keywords go into one list with data on search volume, search intent and competition.
- Keyword strategy: The list becomes the SEO strategy, because every cluster gets a target page and matching content.
- Organic traffic: This creates content that users find in search engines, because it matches their question and purchase intent.
Keyword research for SEA: cost per click and search terms signal purchase intent
Google Ads Keyword Planner: CPC and competition
Google’s Keyword Planner also delivers cost per click and competition per keyword. A high CPC, for example, means that advertisers expect revenue there.
Search terms from running campaigns as a demand source
Running campaigns also deliver real search terms with clicks and conversions. Therefore, the search terms reports flow directly into the demand map.
Align SEO and SEA: budget where organic rankings are still missing
If your website already ranks at the top organically, the need for ads therefore drops. Where rankings are missing, however, Google Ads bridges the time until SEO takes effect.
Prompt research for GEO: capture questions to ChatGPT, Gemini and AI Overviews
Prompt topics with journey stage, intent and persona
SISTRIX AI Topic Research groups AI dialogs into topics with stage, intent and persona. For example, for the German term “Keyword-Recherche”, the tool counts 88 topics, 90% of them in consideration (October 2026).
Prompt set and monitoring of AI visibility
The topics then turn into a prompt set with and without brand names. Monitoring then measures whether ChatGPT, Gemini and AI Overviews mention your company and which sources they cite.
Prompt volume with judgment: keyword data remains the volume backbone
However, no AI provider publishes demand data (SISTRIX, August 2026). Prompt volumes in tools are therefore models built on Google data.
ChatGPT also corresponds to around 12% of Google search volume (Ahrefs, February 2026). Search volume from Google therefore remains the basis of every estimate.
Five levels of demand analysis: search volume, questions, prompts, visibility and buyer questions
Each level answers its own question. Together they give a reliable picture, because the weaknesses of single sources balance each other out.
| Level | Data source | Result |
|---|---|---|
| Search volume | SISTRIX, Google Trends, Google Ads Keyword Planner | Search volume and trend over 24 months |
| Questions | Question keywords, “People also ask” in Google, SEOlyze | Questions per topic for FAQs and guides |
| Prompts | SISTRIX AI Topic Research | Prompt topics with journey stage and persona |
| Visibility | Rankings, SISTRIX AI, Rankscale, Search Console and Bing Webmaster Tools | Current status in Google and in AI answers |
| Buyer questions | Sales, CRM, customer service | Objections and buying criteria from real conversations |
Demand map: the data in every row
Every keyword and every prompt therefore gets the same fields. As a result, topics become comparable and sortable by potential.
| Field | Data source | Content | What it is used for |
|---|---|---|---|
| Search volume and trend | SISTRIX, Google Ads Keyword Planner | monthly search volume, seasonality, development | estimate demand per topic |
| Journey stage | SERP analysis, SISTRIX AI | research through use and service | make keywords and prompts comparable |
| Search intent | SISTRIX, SERP analysis | know, do, visit, website | choose the right page type |
| Cost per click (CPC) | Google Ads Keyword Planner, SISTRIX | CPC and competition in Google Ads | estimate purchase intent and SEA budget |
| Competition and difficulty | SISTRIX, Majestic | ranking domains, page types, link strength | set realistic goals |
| Own position | Google Search Console, SISTRIX | ranking, ranking URL, cannibalization | expand existing pages before writing new ones |
| AI prompts | SISTRIX AI, Rankscale | questions, comparisons, mentioned brands and sources | plan GEO content and monitoring |
| Local relevance | Local Falcon, Google Search Console | “near me” searches, local queries, location pages | align location pages and profiles |
| Revenue value | Growth strategy | demand × click-through rate × lead rate × customer value | priorities by growth per dollar |
From demand to revenue value: data foundation of the growth strategy
The demand per stage then flows directly into the growth strategy. There, demand × click-through rate × lead rate × customer value gives the revenue value per topic.
- Purchase intent first: Do keywords and prompts from selection and purchase therefore get priority.
- Existing before new: We expand pages ranking in positions 5 to 20 first.
- One topic, one page: Clusters also replace single keywords and prevent cannibalization.
- Competitors as the yardstick: The benchmark also identifies which share the competition holds per stage.
Demand analysis process in four steps
digitalagenten runs the demand analysis in four steps:
- We first define topics and markets: services, target groups and locations.
- Keywords and prompts come from SISTRIX, Search Console, Keyword Planner and questions from sales.
- Then every cluster gets a journey stage and target page, plus search intent and revenue value.
- The backlog with briefings follows, then monthly monitoring.
Demand analysis as an ongoing process: every round sharpens the estimate
digitalagenten updates the demand analysis in every round:
- We update rankings, clicks and AI mentions per cluster every month.
- New questions from Search Console, sales and AI prompts then go into the demand map.
- Real clicks and leads gradually replace the assumptions per stage.
Recursive loop
Specialists steer
AI speeds up every round
Layers
Every layer runs through every round.
Demand analysis result: content and prompt backlog with target pages
- Demand map with clusters, journey stages and target pages
- Priorities by revenue value, purchase intent and ranking chance
- Briefings for new and existing pages, including keyword list and questions
- Prompt set for monitoring AI visibility
- Budget recommendation for SEO and Google Ads
Request a demand analysis for your business
In a free initial consultation, we clarify topics, markets and the stages of your customer journey.
Demand analysis, audit and implementation from a single partner
The demand analysis identifies what your target group searches for and what it asks AI. The SEO and GEO audit then checks how much of it your website covers today.
Finally, implementation is handled by our SEO agency, SEA agency, GEO agency and SEO content team.
Demand analysis pays off for these companies
- B2B companies with long decision paths and many stakeholders
- Brands with many locations, where local searches and prompts matter
- Insurers, car dealerships and medical practices, because their customers go through clear stages from need to service
- Marketing teams that plan content based on data and have to justify budgets
- Websites before a relaunch or before entering new markets or languages
FAQ on keyword research and demand analysis
Answers to common questions about keyword research, prompt research and demand analysis.
What is keyword research?
Keyword research identifies which keywords and questions your target group uses in Google and other search engines. It also covers search volume, search intent, competition and the right target page.
What is the difference between keyword research and keyword analysis?
Research collects terms and questions. Analysis then rates the keywords by demand, intent and ranking chance and assigns them to pages.
What is prompt research?
Prompt research captures questions that users ask ChatGPT or Gemini, for example. It also identifies which brands and sources the answers mention.
Is there search volume for prompts?
Reliable numbers hardly exist, because AI providers do not publish demand data. Therefore, we estimate the weight of prompts using keywords from the same journey stage.
Which tools does the demand analysis use?
SISTRIX for search volume, intent and prompt topics, Google Search Console for your own data and the Google Ads Keyword Planner for cost per click. In addition, SEOlyze delivers WDF*IDF and Rankscale tracks AI mentions.
Does keyword research also help with Google Ads?
Yes. Cost per click and search terms from campaigns signal purchase intent. As a result, SEO and Google Ads use the same keyword set.
How long does keyword research take?
The duration depends on topics and markets as well as on languages and locations. After the initial consultation, you receive a binding schedule.
What does a demand analysis cost?
The effort depends on the scope. Therefore, you receive a binding quote after the initial consultation.
How often should the demand analysis be updated?
Immediately for new services, markets or a relaunch, then as monthly monitoring, because search behavior and AI answers change constantly.
How do I do keyword research?
Keyword research starts with a company’s topics and services, collects search terms and questions with tools such as SISTRIX and the Google Keyword Planner and evaluates them by search volume, search intent and competition.
digitalagenten adds prompts for AI assistants and assigns all terms to target pages and customer journey stages.
Related services for more visibility and leads
Book a free, no-obligation initial consultation
In the initial consultation, you receive an analysis and a concrete proposed solution.