Most brands still treat keyword research like a volume report. They export a list from a tool, sort by search volume, pick the biggest numbers they think they can win, and hand the list off. Then they wonder why ranking pages do not convert and why the articles they published never built real authority.

That workflow was already breaking in 2022. In 2026, with AI Overviews appearing on a large share of informational queries and search engines reading entities instead of strings, it does not work at all. Modern keyword research is less about finding big numbers and more about mapping what people actually want, which pages should earn the clicks, and how each keyword fits into a cluster your brand can legitimately own.

This guide covers how we approach keyword research at EmberTribe for DTC brands and growth-stage SaaS companies, and the mistakes we see burning budget on content and paid search.

What Keyword Research Actually Is in 2026

The short version: keyword research is the process of discovering the queries your potential customers type, ask, or prompt, then understanding the intent behind each one well enough to decide what type of asset should answer it.

The old definition stopped at "find search terms with good volume and low difficulty." The new definition has to account for four shifts:

• AI Overviews and zero-click results. Google's AI Overviews now appear on the majority of informational queries, with some studies putting the reduction in clicks on those searches close to 60%. Ranking is no longer the same thing as earning the click.
• Entity-based search. Search engines have moved from matching strings to understanding things. They read your page as a set of entities and relationships, and they assess your site's authority on a topic rather than a single term. This is why topical authority has become central to ranking in 2026.
• SERP features as real estate. People Also Ask, featured snippets, and product carousels now occupy more of the page than the classic ten blue links. The SERP you are actually competing for decides what kind of content wins.
• Generative search as a second audience. Your keywords need to earn clicks in Google and show up as citations in AI answer engines. Those are overlapping, not identical, requirements.

Keyword research that ignores any of these produces the same thing it always did: a spreadsheet with big numbers and no plan.

Intent First, Volume Second

The biggest upgrade you can make to your keyword research process is to lead with intent and treat volume as a tiebreaker, not a filter.

Every query sits in one of four traditional intent buckets: informational, navigational, commercial investigation, or transactional. In 2026, that classification is not granular enough. The pages that win now match narrower sub-intents, things like comparative, instructional, reassurance, and problem-solving intent, each of which calls for a different content format.

A keyword like "best running shoes for flat feet" looks transactional on the surface. Look at the SERP and you see listicles, shoe brand category pages, and a People Also Ask block full of medical questions. The real intent is comparative and reassurance-driven, so a product page will not win that query. A comparison guide built around pain points will.

The practical workflow we use:

1. Pull keyword candidates from your tool of choice.
2. Open the actual SERP for each serious candidate.
3. Tag the intent based on what ranks, not what the keyword sounds like.
4. Group keywords by intent, then by topic. This gives you clusters that map cleanly to content types.
5. Only then look at volume, difficulty, and cost per click.

This is slower than sorting a CSV. It stops you from chasing terms you cannot rank for, and it tells you exactly what kind of page to build.

Ecommerce Keyword Research Is a Different Job

General SEO advice falls apart fast when applied to an ecommerce catalog. Ecommerce brands do not need one keyword per post. They need an architecture that maps collections, products, and content to different layers of demand.

As an ecommerce SEO consultant, the first thing we do with a new DTC client is separate their keyword universe into three jobs:

Collection page queries. These are your category-level commercial terms, things like "merino wool base layers" or "leather crossbody bags." They have the broadest commercial intent and drive the most organic revenue per page. Each collection page should own one primary keyword and three to five secondary terms, with supporting content cleaned up so the collection is the clear canonical answer.

Product page queries. These are the narrower, often long-tail terms that signal a shopper near the bottom of the funnel. "Smartwool 250 base layer men's medium" converts at rates a generic category page cannot touch. Most brands underinvest here because the volume looks small, even though revenue per click is the highest in the catalog.

Informational queries. These are the upper-funnel questions, buying guides, and problem-led searches that feed category pages with topical authority. They rarely convert directly. They exist to help collections rank and to earn citations in AI answer engines. This is where most brands working with ecommerce seo companies fall short: they either skip informational content entirely, or publish it in isolation with no link path to the commercial pages.

The mistake we see most often is treating every keyword as equally valid for any page type. If product pages target the same terms as collection pages, you are competing with yourself. If blog content is not explicitly feeding topical authority into your collections, it is a cost center pretending to be ecommerce content marketing.

How to Cluster Keywords (Without Breaking the Map)

Clustering is where intent work turns into a content plan. A good cluster is a small set of related keywords that share a primary intent and can be answered by one page well enough to compete. A bad cluster is a dumping ground for anything that shares a noun.

Our rule of thumb: if you can write one honest answer that satisfies every keyword in the group without contradicting itself, it is a cluster. If you cannot, split it.

Inside a cluster, one keyword is the anchor. That is the term that drives the URL, the H1, and the canonical intent of the page. The rest are secondary terms you weave into H2s, FAQs, and body copy. This matches how search engines actually read pages in 2026, where entity relationships and semantic context matter more than exact-match keyword density.

From Clusters to Pillars

Across a site, clusters roll up into pillars. A pillar is a broad topic your brand wants to be known for, supported by five to twenty interlinked cluster pages. That is how topical authority gets built, and why one-off posts rarely move rankings anymore.

For a SaaS company, a pillar might be "product-led growth" with clusters for activation metrics, freemium models, onboarding flows, and expansion revenue. We walk through how this shows up in practice inside our complete guide to SaaS SEO, and it is one of the specific things to ask about when you are vetting a SaaS SEO agency.

Keyword Research Serves Content and Paid Search Together

One quiet upgrade modern keyword research makes possible is using the same work to brief both SEO and paid search teams. They are usually treated as separate workstreams with separate keyword lists. That is wasted effort, and it creates inconsistent messaging across the funnel.

The paid team cares about commercial intent, cost per click, and conversion rate. The SEO team cares about volume, difficulty, and topical fit. Intent-tagged, clustered keyword research gives both teams what they need from one source of truth.

A few patterns we use consistently:

• Use paid search as a testing ground for SEO. Terms that convert well in Google Ads are almost always worth pursuing organically. Paid tells you which commercial queries have real buyer intent before you invest months in content.
• Use SEO to lower paid CPCs. When a brand ranks organically for a commercial term, it can reduce bids on that term without losing total traffic, which frees budget for testing new terms.
• Align landing pages across both channels. If paid ads point to a product page and organic traffic lands on a blog post for the same intent, you are sending mixed signals to users and search engines.

One well-built keyword map can inform ad group structure, negative keyword lists, ad copy angles, and a content calendar at the same time.

Tools We Actually Use

There is no shortage of keyword research tools. The honest answer is that most of the best data comes from combining two or three, not from buying the most expensive all-in-one platform.

• Ahrefs and Semrush are the workhorses for keyword discovery, competitor gap analysis, and SERP feature tracking. Pick one, not both.
• Google Keyword Planner is free and still the best source for commercial intent signals tied to actual advertiser demand.
• Google Search Console is underrated. It shows the queries your site already earns impressions for, which is often the fastest path to a cluster worth expanding.
• People Also Ask scrapers like AlsoAsked or AnswerThePublic surface the sub-intents and phrasing real users bring to a topic.

The tool matters less than the workflow. A disciplined researcher with Search Console and the actual SERPs will beat a sloppy operator with a five-figure SaaS stack.

Common Mistakes That Waste Budget

The mistakes we see most often when we audit a brand's keyword strategy:

1. Chasing vanity head terms. High-volume, high-difficulty keywords look appealing on a slide and almost never ship measurable revenue for mid-sized brands.
2. Ignoring SERP intent. Targeting a keyword whose SERP is dominated by Wikipedia, forums, or national publishers is a losing bet no matter what the volume looks like.
3. Publishing one-off posts without a cluster. A single article on a topic is usually invisible. Ten interlinked articles on the same topical territory are how you actually rank.
4. Treating keyword research as a one-time project. SERPs shift, AI Overviews expand their footprint, and new sub-intents emerge every quarter. Keyword research has to be a habit.
5. Skipping customer language. Pulling only what tools suggest misses the phrases your customers actually use in support tickets, reviews, and sales calls, which are often your most valuable keywords.

What to Do Next

If your current keyword research is a spreadsheet sorted by volume, start over. Pull your list, open the SERPs, tag intent, and regroup everything into clusters mapped to the page type that should own each group. That single exercise is usually worth more than a new tool subscription.

From there, decide which two or three clusters your brand can legitimately own in the next two quarters, map them to specific collection pages, product pages, or content hubs, and use the same research to sharpen your paid search targeting. The payoff is a keyword strategy that pulls its weight in both channels instead of living in isolation on a strategist's laptop.

EmberTribe runs keyword research as part of every integrated paid media and SEO engagement for DTC and SaaS clients, which means the work never sits on a shelf. If your keyword strategy feels more like a list than a plan, we can help you rebuild it around intent and clusters that hold up in AI search.