The Google Ads Insights Page in 2026: Demand Forecasts You Can Budget Against

There is a section in every Google Ads account that most advertisers open once, glance at, and never return to: the Google Ads Insights page. The reason is simple — it shows no spend, no CPA, no ROAS, so it never makes it into the daily routine. Yet it holds data that exists nowhere else in the interface: a demand forecast for your category months ahead, search themes that are growing right now, the composition of the audience that actually converts for you, and an automated explanation of why impressions dropped last week.

This is a working guide to that page in 2026: what each card contains, how to read it, where the data misleads you, and — most importantly — how to convert a demand forecast into an actual budget line instead of an interesting chart. It ends with a 20-minute weekly routine and a FAQ.

What the Google Ads Insights page actually is

Insights is an account-level page in the left navigation (Campaigns → Insights; in the refreshed 2026 navigation it also appears under the performance overview block). It is a stack of cards, each answering a different question. Roughly half of them are built on aggregated Google Search data for your category and geography — not on your account. That is the crucial difference from the rest of your reporting: every other report tells you what already happened to your money, while Insights tells you what is happening in the market around you.

Three reasons it gets skipped:

• It is not denominated in money. No cost column means it does not look like a working tool.
• The card set is unstable. Google ships and retires blocks; availability varies by country, vertical and campaign type. Open it once, see two empty cards, and it is easy to conclude the page is useless.
• The data needs interpretation. “Category demand will rise 35% by mid-November” is not an action. It becomes one only after you run it through your own unit economics.

That last step is the real work, and it is what the rest of this article covers.

What sits on the Insights page in 2026

The exact card set depends on your vertical, country, data volume and campaign types. These are the blocks you will most often see, and what each one is genuinely good for.

Card	Built on	Practical value
Demand forecasts	Aggregated Google Search demand for your category and geo	High: budget and inventory planning
Search trends	Movement in the queries you already appear for	High: finds rising sub-niches before competitors
Consumer interest	Category-level interest in your geo	Medium: context rather than a direct action
Audience insights	Your converting users, mapped to Google segments	High: hypotheses for observation audiences and creative
Performance shifts / explanations	Your stats plus automated decomposition	High: saves an hour of manual diagnosis
Asset and combination insights	Your ads and assets	Medium: direction for RSA testing
Optimisation and recommendations	Google automation	Low to medium: apply selectively

The mental model is straightforward. Market-data cards (forecasts, trends, interest) answer “where is demand going”. Account-data cards (audiences, performance shifts, assets) answer “what happened in my account”. The first group belongs to planning weeks and months ahead; the second to diagnosing yesterday’s drop.

Demand forecasts: reading them and turning them into budget

The demand forecast card projects how Google expects search interest in your category to move in a specific country over the coming months — typically a three-to-six-month horizon with weekly granularity. It is not a forecast of your sales and not a forecast of your clicks. It forecasts category demand.

What you are actually looking at

• Category. Assigned automatically from your keywords, landing pages and product feed. Check this first: if Google has filed you in the wrong bucket — common for narrow B2B and multi-category retailers — the entire forecast is irrelevant to you.
• Geography. Usually country level. If you advertise in one region inside a country, the forecast will be blurred.
• Peak. The expected peak week or date range, with a percentage lift against the current baseline.

The formula that turns a forecast into a number

The forecast speaks in demand; you need money. The bridge looks like this:

Peak-week budget = current weekly spend × (1 + forecast demand lift) × auction pressure factor

The auction pressure factor accounts for competitors entering the auction at the peak and driving CPCs up. In competitive commercial verticals it typically lands somewhere between 1.1 and 1.4 — treat that as a benchmark to sanity-check against, not a constant. Calculate your own from historical CPC during previous peaks using a day-level report. In low-competition B2B niches it can sit close to 1.0.

Worked example: you currently spend $3,500 per week. The forecast shows a 40% demand lift for the week of 16–22 November. Your historical November CPC inflation is around 20%, so your factor is 1.2.

3,500 × 1.4 × 1.2 ≈ $5,880 — that is the peak-week budget required to hold your current impression share.

Then comes the reality check: can your allowable CPA absorb a 20% CPC increase? If not, the correct answer is not “spend more” but “hold budget through the peak and let impression share go”. How to derive an allowable CPA from margin and lifetime value is covered in the guide to media buying unit economics and max CPA from LTV.

Where the demand forecast gets it wrong

• Narrow niches. The tighter the category, the wider the confidence interval. For specialised B2B the card is often simply absent — which is more honest than an invented number.
• New categories. Forecasts are built on history. For a product category that did not exist a year ago, the card has nothing to work with.
• Event-driven spikes. Regulatory news, a viral moment, a competitor’s launch — none of it is predictable from historical seasonality.
• Category demand is not your demand. A 40% category lift does not mean 40% more conversions for you. If competitors add budget more aggressively, you can see more impressions and a falling impression share at the same time.

So the forecast is an input to planning, not its output. Cross-check it against Performance Planner for defensible budgets, which models your own spend-to-conversion curve, and against bid and budget simulators at campaign level.

Search trends: finding rising sub-niches early

The search trends card shows which queries and query categories in your space are growing or shrinking over the selected window. Unlike the Google Ads Keyword Planner, which returns averaged volumes and forecasts for specific phrases, trends show movement — what is accelerating right now.

A workable process:

1. Set the window to 28 or 90 days. Seven days is mostly noise.
2. Sort by percentage growth, but immediately discard themes with a trivial base. A 900% lift on twelve impressions means nothing.
3. Write down five to ten growing themes you can genuinely serve with your product.
4. For each, check whether a matching landing page exists. If it does not, the theme becomes a content task, not a campaign.
5. Growing themes you cannot serve go straight into negative keywords before automation starts spending on them. The mechanics are in the guide to negative keywords and search terms.

The underrated value of this card is early warning on seasonal shifts. If a theme that normally accelerates in October started climbing in late August, that is a reason to move your campaign calendar rather than wait for the date.

Audience insights: who actually converts for you

The audience card takes your converting users and maps them onto Google’s segments — affinities, in-market intent, demographics, life events. It answers a question no standard report can: what else are the people who buy from you interested in?

What to do with it:

• Do not narrow targeting immediately. The temptation is obvious: 60% of buyers index into “fitness enthusiasts”, so target that segment. The correct first move is to add the segment in observation mode and read the data without cutting reach. The mechanics are in audience observation without narrowing reach.
• Use it as a creative brief. If your buyers overindex on “recently moved”, that is a ready-made angle for an RSA headline.
• Feed PMax and Demand Gen. Discovered segments become audience signals in Performance Max and raw material for custom intent segments. The next step in the same direction is GA4 predictive audiences, where the segment is built from purchase probability rather than declared interest.

The constraint is volume. At 20–30 conversions a month the card is either empty or statistically meaningless. As a rough threshold, treat it as usable from roughly 100–150 conversions in the analysis window.

Performance shifts: diagnosing a drop in ten minutes

The performance-shift card (labelled Explanations in some interface versions) decomposes a change in spend, clicks or conversions into contributing factors: bid changes, budget changes, auction volume, CTR, query mix.

In practice, when conversions are down 30% on a Monday morning, open this card before spending an hour digging manually. Then verify what it tells you:

1. Card says “auction volume decreased” → cross-check impression share and auction insights. If share held steady while impressions fell, the market moved, not you.
2. Card says “bid or budget change” → go to change history and find the exact edit and who made it.
3. Card says nothing while conversions collapsed on stable clicks → this is almost always tracking. Follow the sequence in the 30-minute conversion tracking diagnosis. The mirror-image problem — conversions suddenly reading higher than reality — is covered in the guide to duplicate conversions and deduplication.

One caveat worth internalising: the card explains correlation, not cause. It will tell you CTR fell 18%. It will not tell you a competitor launched a stronger offer.

Five things the Insights page cannot do

• It does not do money. No CPA, no ROAS, no margin. All financial modelling stays on your side.
• It is not competitive intelligence. You will not see who entered the auction or with what offer.
• It needs volume. On an account doing 500 clicks a month, half the cards will be blank.
• It does not export well. There is no API access to demand forecasts. If you want history, record the numbers weekly into your own BigQuery data warehouse.
• It cannot separate your growth from the market’s. Incrementality is not measured here; that requires geo-experiments and incrementality testing.

A worked quarter: from card to plan

Abstract advice transfers badly, so here is an end-to-end example. A mid-sized home appliance retailer, single country, $3,500 weekly spend, average CPA $28, allowable CPA from margin $34.

Step 1 — forecast. The card shows a category peak in the week of 23–29 November, +55% against the current baseline, with the ramp starting in the first week of November.

Step 2 — auction pressure. Pulling last November day by day, peak-week CPC ran 26% above October. Factor: 1.26.

Step 3 — budget. 3,500 × 1.55 × 1.26 ≈ $6,835 for the peak week, reached in steps: +25% in week one of November, +25% in week two, then the full figure. Jumping from $3,500 to $6,800 overnight would push the strategy back into a Smart Bidding learning period at the worst possible moment.

Step 4 — economics check. If CPC rises 26% and landing page conversion rate holds, CPA drifts from $28 toward $35 — past the allowable line. So before the peak you either lift landing page conversion, lift average order value, or consciously concede impression share and protect CPA. Conceding is a legitimate decision — it is just far better made in September than discovered in December.

Step 5 — trends. Over 90 days, “self-emptying robot vacuum” is up 64% and “clothes steamer” is up 41%. The first is in your assortment: it becomes its own ad group with dedicated headlines. The second is not: it goes into negative keywords so broad match does not buy it at peak-season CPCs.

Step 6 — audiences. Among converters, the “recently moved” life event overindexes. It does not narrow targeting; it becomes an audience signal for PMax and an angle for one RSA variant.

Four cards, four concrete actions, one management decision about impression share. That is the page working as intended.

Common mistakes

• Treating a forecast as a commitment. It is a range, not a promise. Plan a base case and a case 30% below it.
• Loading budget a week before the peak. Smart Bidding needs adaptation time. Step budget up 20–30% every three to four days instead.
• Narrowing targeting off a single audience card. Correlation between a segment and conversions does not mean conversions do not exist outside it.
• Ignoring empty cards. An empty card is information too: usually it means the account lacks statistical volume, and that is the thing to fix first.
• Chasing every trend. Of ten growing themes, two are relevant to your product. The other eight are a way to dilute budget.

The Insights page does not tell you what to do. It tells you what the market is doing before it shows up in your CPA. The value appears the moment a demand forecast becomes a budget line and a rising theme becomes a landing page task.

Where to go from here

If your account is new and most cards are blank, the problem is not the report — you lack history and volume. In that situation it makes more sense to build a reliable signal first (conversions, values, first-party data) and return to Insights later. For teams that need an account with existing history and lifted limits from day one, PPC Rebels offers agency Google Ads accounts for high-volume advertisers. If the bottleneck is methodology rather than access, start with a structured Google Ads account audit.

Related reading: DSA to AI Max Migration: The 2026 Playbook for Search Campaigns

FAQ: the Google Ads Insights page

Where do I find the Insights page in Google Ads?

In the left-hand navigation at account level: Campaigns → Insights. In the 2026 navigation it is also reachable from the performance overview block. It exists at account level rather than per campaign.

Why are half of my cards empty?

Most often insufficient data. Account-data cards need conversion volume; market-data cards need a recognised category and a supported country. Some blocks are also unavailable for certain campaign types and markets.

How accurate are Google Ads demand forecasts?

Accuracy improves with category breadth and stability. In large seasonal categories — electronics, travel, gifting — direction and peak timing are usually right while amplitude carries error. In narrow or new categories, treat the forecast as a weak signal. Always plan a base and a conservative scenario.

How is a demand forecast different from Performance Planner?

The demand forecast models the market: how many people will search your category. Performance Planner models your account: how many conversions you get at a given spend. The first answers “will there be demand”, the second “what will it cost me”. Use them in that order.

Can I pull Insights data through the Google Ads API?

There is no full API access to demand forecasts and trend cards. Some data can be exported in the interface, but without historical depth. If you need a time series, record the figures manually on a weekly cadence into your own storage.

Google assigned my account to the wrong category. Can I change it?

There is no direct category editor for demand forecasts. The assignment is inferred from keywords, landing page content and product feed. A clearly wrong category is a signal that your account structure or landing pages give Google an ambiguous picture of the business — that is where to start.

Can I use audience insights with low conversion volume?

Below roughly a hundred conversions in the window, no. A more productive path at low volume is to add micro-conversions to build signal, then return to audience analysis once the numbers support it.

Should I apply the recommendations shown there?

Selectively. Some are genuinely useful — fixing tracking errors, adding missing assets. Others exist to increase spend. Never enable auto-apply on a budget-constrained account. A breakdown of what to apply and what to dismiss lives in the optimisation score guide.

Can I plan a budget from the demand forecast alone?

No. The forecast knows nothing about your margin, site conversion rate or allowable CPA. A budget emerges at the intersection of the forecast, your unit economics, and Performance Planner output.

How often should I check the Insights page?

Weekly for trends and performance shifts; monthly for audience insights. Daily checking serves no purpose — the data does not refresh that fast, and reacting to weekly noise produces unnecessary edits and resets algorithm learning.

Does the Insights page help with competitor analysis?

Only indirectly. It will tell you auction volume grew or a theme is gaining traction, but it will not name competitors or show their offers. For that you need auction insights and external SERP tooling.