Skip to content
Dmitry FabushevHEAD OF SEO
← All cases/SEO operations
Operating in production

SEO Operating Model: roadmap, monitoring and business demand in one system

In a five-person SEO team, roadmap work, recurring operations, monitoring signals and business requests compete for the same execution capacity. I used the move from Jira to Kaiten not as a tool migration, but as an opportunity to redesign the operating model, automate repeatable work, connect monitoring to execution and formalize unplanned demand.

Case at a glance
What was at stake

Without explicit rules, roadmap, BAU, monitoring response and business demand compete for the same capacity invisibly. Priorities become harder to read, recurring work depends more on manual reminders, and unplanned demand can displace planned work without an explicit decision.

What I owned

Designed the SEO operating model: intake, ownership, capacity rules, automation, recurring work and the cross-functional demand model.

Outcome

Executable SEO work now runs through one managed operating layer: every task has a visible Kanban stage, recurring work is created automatically, meaningful monitoring signals get an owner, and demand from roughly 50 category managers enters through one channel with a quota of up to 10 categories per month.

Four work sources → one managed flow
Where work comes from
RoadmapRecurring workMonitoring signalsBusiness demand
How work moves through the system
01Intake02Priority03Owner04Delivery05Measurement
Capacity protection
~50 category managers1 channelup to 10 categories / month

Roadmap, recurring work, monitoring and business demand share one workflow; the quota protects planned team capacity.

Why this matters

The outcome is managed capacity, not a new task board

Roadmap work, recurring operations, monitoring response and business demand now enter one operating layer. The result is visibility not only into the backlog but into where team capacity is being consumed, with consistent ownership and execution rules for specialists.

Key decision
02

Do not ban unplanned demand — make it limited and transparent

Commercial demand cannot be removed: assortment, seasonality and business priorities move faster than the roadmap. But it should not consume team capacity invisibly.

Before

Roughly 50 category managers could potentially create parallel incoming demand whose total load was difficult to see as a separate capacity stream.

After

Requests go through one responsible channel. Unplanned SEO optimization is limited to up to 10 categories per month, so the business keeps fast access to SEO while roadmap capacity is protected.

01

Why the task tracker was not the real constraint

Jira itself was not the core problem. Four sources of work existed at the same time: roadmap initiatives, recurring SEO operations, monitoring-triggered work and business requests. The move to Kaiten became an opportunity to define one intake, ownership and prioritization model instead of reproducing the old queue of cards.

03

How a monitoring signal becomes execution

An analytical signal creates value only when it gets an owner and ends in a decision.

01

Signal

Monitoring detects a meaningful deterioration in a cluster.

02

Owner

The responsible specialist automatically receives a task to investigate that specific cluster.

03

Validation

The specialist confirms or rejects the signal, checking SERP rotation, parser collection issues and other false-positive causes.

04

Evidence + hypothesis

If the change is real, supporting data is collected and a testable explanation is formed.

05

Action

A supported hypothesis becomes a change or a separate SEO task.

06

Measurement

After implementation, the cluster remains under observation to verify the actual result.

04

What was standardized and automated

Automation removes predictable administrative work while leaving decisions with specialists.

~20 automations

Workflow actions, small utility scenarios, corporate-email webhooks and automatic creation of selected tasks.

~10 recurring tasks / month

Recurring work appears on schedule and depends less on an individual remembering to initiate it.

5 task types + 12 checklists

Standard processes start with the same structure, context and control points.

05

Cross-functional operating layer

The operating model has to work beyond the SEO team because part of delivery depends on adjacent functions.

Content management

Dependencies and shared work move through an explicit workflow instead of scattered messages.

Paid promotion

Related search work can be coordinated with clear ownership on both sides.

Category managers

Roughly 50 category managers use one responsible representative and a quota-based intake path.

Product / Development

Work requiring product or engineering changes stays connected to the originating SEO priority and downstream delivery.

06

My responsibility vs the team

Designing the operating model and doing the daily work inside it are different layers of responsibility.

My responsibility

Operating model, work classes, intake, ownership, capacity rules, automations, recurring workflows, templates, integrations and cross-functional demand.

SEO team

Execute work, validate signals, collect evidence, form and test hypotheses, implement changes and monitor outcomes.

Context & measurement
Context
Large e-commerce business · SEO team and cross-functional operations
Scale / period
5-person team · ~40 planned tasks/month · up to ~15 operational items per specialist · ~50 category managers → 1 channel

How to read the workload

~40 refers to planned tasks across the team per month. In addition, each of the four specialists can receive up to ~15 operational items depending on monitoring signals and incoming demand. An operational item may be a signal check or standard procedure and is not necessarily a standalone large project.

SEO operations

Facing a similar challenge?

Tell me where SEO is blocked by product, engineering or data.

DiscussRelated case: 27,000 queries every day: from monitoring to owned SEO actionsOther cases