The role of an SEO consultant in 2026: work and deliverables

An SEO consultant is judged less on what they know than on what they move forward in your organization, week after week. This article describes an engagement as it actually runs: how it is framed, the pace it runs at, what it produces, who has to be mobilized on your side, and what is committed to as opposed to what cannot be guaranteed. If your question is rather about which support model to choose, that framing belongs with the choice of an SEO, GEO and link building agency.

What an SEO consultant does, and what they do not do

The SEO consultant works to improve a site's organic visibility and, increasingly, its presence in generative environments (engines with AI overviews, LLMs). The role is pragmatic: turn findings into executable actions, then steer their impact over time. The main expectation on the company side is a clear execution plan — “what to do, where, in what order, against which validation criteria” — rather than a list of best practices. That is what separates an engagement that moves from an engagement that produces documents.

Consultant, expert, specialist: three roles that get confused

Organizations often confuse three roles that are not bought for the same reasons:

• The SEO consultant: they move the project forward through a method, rituals, deliverables and cross-team coordination.
• The expert: they step in on high-level trade-offs, complex cases and structural decisions (redesign, international, architecture issues).
• The specialist: they concentrate on a vertical scope (technical, content, local, international and so on) or on one type of site.

These are neither three levels of seniority nor three price brackets: they are three ways of occupying the subject, and the same person can combine them depending on the phase. The useful question at scoping: do you expect someone to settle a few hard points in your place, or to hold execution over time?

What stays with other functions

A scope is also described by its limit. The consultant specifies, prioritizes, runs the acceptance check and measures; they do not replace the functions that produce.

• Content production at scale stays with an editorial team or with writers: the consultant supplies the brief, the intent and the review criteria, not the volume.
• Development and release stay with the technical team: the consultant writes the ticket, the global rule and its exceptions, then verifies after go-live.
• Design and usability stay with the product teams: the consultant flags what gets in the way of crawling or conversion, they do not redraw the templates.
• Media buying belongs to another discipline: it is coordinated with organic, it is not steered from the SEO engagement.

Setting that boundary at scoping avoids the classic disappointment: a sound roadmap that nobody is responsible for executing.

A typical week and the steering rituals

A “realistic” week often splits into four blocks that run in a loop:

• Analyse: Search Console and Analytics, crawl, indexing, pages that stall, CTR variations, technical anomalies.
• Prioritize: turn a list of topics into a maximum of 10 ordered actions, with owners and acceptance criteria.
• Execute: tickets for the technical team, briefs for content, internal linking work, template optimization, approval before release.
• Measure and iterate: check indexing, ranking gains, the impact on clicks and conversions, then adjust.

The key point: “SEO” output is only worth something if the organization can absorb the recommendations (development, content, approval, publication). So the consultant spends as much time making things executable as analysing.

Three rituals are enough to hold that pace: a weekly check-in of 30 to 45 minutes on ticket progress, trade-offs and blockers; a content review, weekly or fortnightly, which approves the briefs and checks internal linking and templates; a monthly KPI review on trends, what worked and what should be stopped. Good steering means every action carries an owner, a date, a definition of “done”, and an expected indicator (even an imperfect one). It is that last point, the one most often forgotten, that turns a monthly review into a decision meeting.

The core does not change from one site to another — indexing, content, authority, measurement — but the day-to-day depends on the model: in B2B, long cycles and fine-grained conversion tracking; on a brochure site, trade-offs on a few strategic pages and gains tied to CTR; in e-commerce, volume, facets, URL duplication, pagination and templates; in media, publishing cadence, cannibalization and structuring into hubs.

Who does what: what the engagement asks of your teams

Coordination is not free: every counterpart has to supply something in return, otherwise the recommendations stay in the backlog. That split is settled at scoping, not at the first blocker.

From scoping to execution: the working method

Scoping avoids the “we do SEO” trap with no compass. It documents the objectives (leads, MQLs, demos, revenue, share of voice), the scope (domains, directories, countries, page types), the constraints (CMS, release cycles, legal approval, available resources) and the KPIs — including, in 2026, exposure indicators in generative answers.

Then comes the baseline, without which no impact can be proved. Two tools are enough: Search Console for what happens before the click (queries, impressions, CTR, position, indexing), Analytics for what happens after (engagement, conversions, value). To that is added a reproducible technical picture: crawl, HTTP statuses, indexability, canonicals, depth, internal linking. This step also spots the blocking workstreams — crawling, indexing, 4XX and 5XX errors, duplication — before investing in content production.

Map intents, then prioritize by impact, effort and risk

Opportunity work is not just a keyword list: it is a mapping of intent → page type → target page. The consultant groups intents (informational, commercial, transactional, navigational), assigns each one a suitable format (guide, service page, category, FAQ, local page), and avoids cannibalization by defining one “reference” page per intent.

SEO backlogs being infinite, prioritization connects three axes: impact (crawling, indexing, CTR, positions close to page 1, conversions), effort (complexity, dependencies, production lead time) and risk (regression, side effects, uncertainty on templates).

A useful benchmark for framing impact: according to our SEO statistics, the click-through rate on the first organic position reaches 34% on desktop (SEO.com, 2026), while page 2 falls to 0.78% of clicks (Ahrefs, 2025). Hence the value of gaining a few places on queries already close to the top 10.

Tickets, acceptance criteria and technical checks

The difference between an engagement that moves and one that stalls often comes down to the “recommendation → execution” conversion. Three artefacts carry it, and none replaces the others:

• tickets (context, problem, solution, pages affected, global rule vs exceptions);
• acceptance criteria (how to check it is correct, and by whom);
• editorial briefs (intent, structure, evidence, internal linking, brand constraints).

The on-page routine that follows: revising page structures, aligning intent and content, organizing internal linking, standardizing templates when volume demands it. On the technical side, three checks come before the others, because nothing pays off until they are settled: indexability (robots.txt, sitemap, directives, canonicals), HTTP statuses (limiting 404s and 5XXs, avoiding redirect chains), performance and the mobile experience. On that last point, the argument that gets a performance backlog through a steering committee is not a technical one: 40 to 53% of users leave a site if loading is too slow (Google, 2025).

The deliverables of an engagement: what you receive

Four deliverables come up in almost every engagement. What sets them apart is not their name — everyone delivers an “audit” — but their content, and the evidence that they can actually be used.

Reporting usually feeds a monthly “CTR batch”: titles, descriptions, richer snippets. It is the least expensive workstream of the engagement — an optimized meta description is associated with +43% CTR (MyLittleBigWeb, 2026), an order of magnitude that varies by query: you recover clicks on impressions you already have.

Documentation and handover: the most underrated deliverable

Documentation prevents dependence on a single person. It is produced as the engagement goes, not at the end: a handover written in the last two weeks is a failed handover. What is documented, at minimum:

• the rules on URLs, canonicals, pagination and facets;
• the acceptance check-list before publication;
• the content update process (refresh);
• the internal linking conventions (hubs, anchors, “pillar” pages).

The aim: stop a departure or an absence from freezing SEO in the backlog. It is also a test of good faith — whoever documents accepts becoming replaceable.

What GEO adds to the engagement

In 2026, performance is no longer summed up by positions: engines display summaries and part of the usage is shifting to generative search. Our GEO statistics record a fall in organic traffic linked to the arrival of generative AI, estimated at between -15 and -35% (SEO.com, 2026; Squid Impact, 2025), but also a rise in impressions of +49% following the launch of AI Overviews (Squid Impact, 2024). The two movements coexist, and that is what forces a rethink of the KPIs. GEO adds one operational objective: becoming citable, being taken up as a source and not only ranked well. It is not a separate workstream, but one more constraint on the content the engagement already produces.

Structuring extractable content, and measuring exposure

To encourage pick-up by generative systems, the consultant favours readable structures: short definitions up front, lists and steps rather than compact blocks, sourced statistics with their year, an FAQ at the end of the page. Pages structured with an H1-H2-H3 hierarchy are x2.8 more likely to be cited by generative engines (State of AI Search, 2025).

Measurement, on the other hand, remains more imperfect than classic SEO measurement, and that has to be said before turning it into a board-level KPI. A pragmatic approach: keep Search Console and Analytics as the base, add “click-free” visibility monitoring (mention frequency, share of voice, presence on the queries that trigger AI overviews), and document the “reference” pages with their evidence.

In-house, agency or training consultant: which collaboration setup

The same work is not steered the same way depending on the execution capacity you have.

• In-house: the issue is continuity — rituals, documentation, shared responsibility. The consultant is often attached to marketing but works across content and technical alike; both doors therefore have to be opened to them.
• With an agency: what changes is capacity (several profiles that can be mobilized) and governance (counterparts, processes). The requirement is unchanged: actionable, prioritized deliverables, and a ticket → production → acceptance check → measurement loop.
• With a training consultant: useful when the bottleneck is execution on the team side. The right balance is to train on real cases (your pages, your data), to produce templates (briefs, check-lists, acceptance checks) and to maintain a delivery cadence — otherwise the training becomes a brake.

The switching condition is easy to state. When the need is multi-profile execution capacity — technical, content, link building and steering in parallel — one person will not be enough, however good they are. When it is a very targeted one-off reinforcement on an identified hard point, a short assignment costs less in coordination. And when the need is team autonomy, the training model is the answer — provided it produces artefacts, not sessions.

Pricing, duration and timelines: what is committed to, what cannot be guaranteed

You are also choosing a billing model, and it has to match the nature of the work expected.

• Monthly retainer: suited to the run — steering, continuous production, iteration, reporting. The scope is described in advance and the workload is smoothed out.
• Time and materials, or a day rate: suited to intensive phases — redesign, migration, in-depth audit, recovery after a drop. You pay for time and keep control of priorities, at the cost of less budget predictability.
• Targeted assignment: useful when you already know the hard point (indexing, templates, editorial plan, internal linking). The deliverable is bounded and the engagement has a stated end.

What makes the budget vary, whatever the model: the size of the site, the technical state at the start, the volume of content to produce, and the coordination load.

How long an engagement lasts, and realistic timelines

Duration depends on context, but three cycles recur: scoping and baseline in 1 to 3 weeks depending on access, volume and complexity; an implementation phase of 2 to 12 weeks, often the least predictable one because it depends on the teams; and finally result cycles, with first trends within a few weeks and structural effects over several months. On results, three horizons stand apart, and confusing them is the most frequent source of disappointment:

• Very short term: fewer critical errors, indexing stabilized, CTR improvements through titles and snippets.
• 1–3 months: impact on the optimized pages (queries close to the top 10), better semantic coverage, more effective internal linking.
• 3–6 months and beyond: more visible effects on content clusters, on authority, on link building and on conversions.

What is committed to, and the promises that should raise a flag

The distinction to hold in a negotiation is clear-cut. Deliverables and deadlines are committed to: an audit by a given date, a prioritized roadmap, a number of briefs per month, monthly reporting. The expected gain is stated as a hypothesis with its conditions attached, then measured. Only a precise position on a precise keyword on a precise date cannot be guaranteed. Three signals should raise a flag:

• promising a “guaranteed” position 1 without mentioning constraints, competition or execution;
• delivering an “inventory” audit with no prioritization, no owners and no acceptance criteria;
• offering risky shortcuts, especially in link building, without explaining the trade-offs and the governance.

Conversely, a provider who refuses to estimate a gain before diagnosis is not a bad sign. What should worry you is a refusal to commit to deliverables and deadlines.

Depending on your situation, the useful detour is not the same:

• Merchant catalogue: if the number of URLs exceeds what can be handled page by page, the indexing framework becomes the subject — facets, variants, seasonality → what an e-commerce SEO consultant handles.
• Budget already committed to paid: if nobody knows what buying keywords adds on top of organic, the trade-off is made query by query → how an SEO and SEA consultant settles it.
• Two candidates to choose between: if you have to judge a real level rather than a stated method, it is the evidence, the reference frameworks and the indicators that settle it → assessing SEO expertise.

Finally, when the engagement goes beyond what one person can carry and several profiles have to work in parallel, what you need to buy is team capacity, not extra days: that is the purpose of SEO, GEO and link building support.

FAQ: choosing and steering SEO support

What are the most frequent day-to-day tasks?

Search Console and Analytics analysis, prioritization, creating technical tickets, writing briefs, optimizing content and internal linking, checking indexing, reporting, and steering rituals with the teams. Time is split roughly equally between analysing and making things executable.

Which standard deliverables should you expect: audit, roadmap and reporting?

An audit with findings, evidence and quick wins; a prioritized roadmap (batches, dependencies, owners, milestones); structured content briefs; monthly reporting geared to decisions, not only charts. And documentation kept up as you go, which secures continuity.

Which tools are essential day to day?

A minimal, robust base is enough: Search Console for the pre-click side (queries, impressions, indexing) and Analytics for the post-click side (journeys, conversions). Beyond that, a crawler for the technical picture and, if the workload justifies it, somewhere to centralize briefs, planning and follow-up.

How long does an engagement last, on average?

You often see 1 to 3 weeks of scoping and baseline, then a run phase of several months. The first signals (CTR, indexing) can move fast, but structural gains consolidate over several months, and the implementation phase remains the least predictable because it depends on your teams.

How do you organize onboarding in-house or with an agency?

Define a steering channel (the weekly check-in), a shared prioritization rule, a single format for tickets and briefs, and an acceptance-check and measurement loop. Also name who approves a brief, who releases and who runs the acceptance check. Without governance, recommendations end up in the backlog.

How do you check the quality of the work, without jargon?

Ask four questions of every recommendation: “What evidence supports this recommendation? What priority, and why? Who executes it? How is it validated?” Add a fifth one for measurement: which indicator should move, and by when?

Should you choose a freelancer or an agency, depending on the context?

A freelancer works if the scope is clear and can be executed with your teams. A firm is useful if several skills have to be mobilized in parallel and continuity has to be secured. In both cases, quality is judged on actionable deliverables and on governance, not on the size of the provider.

What budget should you plan for, given the pricing and collaboration models?

The budget depends first on the model chosen — monthly retainer, time and materials or day rate, targeted assignment — then on four variables: the size of the site, the technical state at the start, the volume of content to produce and the coordination load. Ask the proposal to state which of those variables carries most of the amount: that is where the room for discussion is.

What is the difference between a consultant, an expert and a specialist?

The consultant moves execution forward: method, coordination, deliverables. The expert works more on complex, structural trade-offs (redesign, international, architecture). The specialist concentrates on a sub-area — technical, content, local. The three stances can be combined depending on the company's SEO maturity.

Can SEO and GEO be handled in the same engagement?

Yes, and it is increasingly expected. The SEO foundation — indexing, content, authority — remains the base; GEO adds structuring requirements (extractable content, evidence, sourced data) and new exposure indicators in AI answers, whose measurement remains less reliable than that of classic SEO.

When does a training consultant become useful?

When your main brake is team autonomy — content, development, approval — and not strategy. It works if it produces concrete artefacts (templates, check-lists, examples) without breaking the production cadence. If all that is left after they go is slides, the model has not worked.

Which KPIs should you track to connect SEO, GEO and ROI in B2B?

On the SEO side: impressions, clicks, CTR, positions and indexing in Search Console, plus conversions and their value in Analytics. On the GEO side: exposure indicators (mentions, citations, share of voice) and the quality of traffic from AI environments when it is measurable. On the ROI side: connecting costs (time, production, fixes) to gains over a consistent period.