Content strategy is one of the few agency services where a buyer can pay a substantial fee and receive something genuinely valuable, or pay the same fee and receive a slide deck that restates what the client already said. The difference is rarely visible in the proposal, because both look like a discovery phase followed by a framework. What separates them is whether the engagement ends with decisions that constrain future work, or with a document everyone praises and nobody consults. This page describes what a strategy engagement should actually produce, how it differs from production, and the questions that tell you which of the two you are being sold.
Strategy and production are different purchases
Strategy decides what to publish, for whom, in what order, and how success will be recognised. Production writes, edits, designs and ships it. Many agencies sell both, which is convenient and creates a mild conflict, since the firm recommending how much content you need is often the firm that would produce it. That is not a reason to avoid bundled arrangements, it is a reason to price them as two things and to ask what the strategy would say if you took the production elsewhere. A strategy engagement that concludes you need a large, ongoing volume of content, produced conveniently by the same agency, should be read carefully. So should one that concludes you need almost nothing, from a firm that only sells strategy. Buyers who separate the decisions from the hands usually find that the underlying content strategy services they actually need are smaller and more specific than the first proposal suggested.
What a strategy engagement should leave behind
Five artefacts, and a buyer can ask for them by name. An audit of what already exists, saying plainly what to keep, improve, merge or delete, because most established sites have more pages than they can maintain and consolidation often outperforms creation. A defined audience with the questions each segment actually asks, sourced from search data, sales calls and support tickets rather than from a workshop of assumptions. A prioritised topic plan that says what comes first and why, with the reasoning visible so a new hire can extend it. A publishing model naming who writes, who reviews, who approves and at what pace, which is the part that determines whether anything ships. And a measurement definition agreed before publishing starts. A deliverable missing the third and fourth of those is a research document, and it will not survive contact with a busy quarter.
The quality bar the work is actually judged against
Google's guidance on creating helpful, reliable, people first content is the clearest public statement of what a page has to be, and it reads less like an algorithm description than a description of an honest publisher. Does the content demonstrate first hand experience or genuine expertise. Would a reader feel they got what they came for. Is it written for people rather than assembled to catch search traffic. Google's spam policies are the other half, naming scaled content abuse, meaning pages produced in volume mainly to rank rather than to help. Those two documents together explain why volume led content programmes have become steadily worse investments: producing many mediocre pages is now cheap for everyone, so it is no longer a competitive advantage. Ask a candidate agency how it decides a page is worth publishing at all, and whether it has ever recommended a client publish less. The answer separates a strategist from a production shop with a strategy phase attached.
How to buy it without overpaying
Start bounded. A defined strategy project with a stated deliverable and an end date tests how a firm thinks for a known cost, and it produces something you own regardless of whether the relationship continues. Ask to see a redacted strategy document from a previous client before committing, and read it as if you had to act on it tomorrow: is it a set of decisions, or an elegant description of your situation. Ask how much of your team's time the process assumes, since strategy work stalls on stakeholder availability more often than on agency capacity. Agree that everything produced belongs to you, including the research behind it. And agree the review point: a strategy is judged by whether the plan was followed and whether following it produced the intended result, which means a fair assessment sits at least two quarters after the first pages ship, not at the moment the document is delivered.
Questions people ask about content strategy agency
How much should a content strategy cost?
It scales with the size of the estate being audited and the number of stakeholders to be interviewed, so a single product company and a large organisation with several business units are not comparable purchases. Rather than benchmarking a figure, get two or three firms to quote the same written scope, which makes the differences readable.
Do we need a strategy if we already publish regularly?
Often yes, and the value is usually in subtraction. Established publishers commonly have overlapping pages competing with each other, outdated pieces still attracting visitors, and a topic focus that drifted. An audit that identifies what to consolidate or retire frequently returns more than a plan for new work, and it is cheaper.
Should the same agency write the content?
It can be efficient, and the conflict is worth naming out loud. The firm recommending volume is the firm that would be paid for volume. Ask what the strategy would recommend if production went elsewhere, price the two separately, and keep the option to split them. A confident strategist will answer that question without discomfort.
How do we measure whether the strategy worked?
Agree the definition before publishing. Usually a small set: qualified enquiries attributable to content, growth in visibility for the topics you chose to own, and whether the publishing model actually held. Page counts and traffic totals are inputs. If the plan shipped on schedule and moved none of the agreed measures, that is a clear answer and a useful one.