Marketing
A digital marketing agency's first 90 days: a practical client plan
An example 90-day marketing plan covering baseline measurement, customer interviews, landing pages, content experiments and review decisions.
Use a plan, not a revenue promise
The first 90 days of an agency engagement should establish a reliable baseline, improve one important customer journey and test a small number of distribution ideas. The schedule below is a suggested operating plan, not a claim that every business will see growth within three months. Sales cycles, existing demand, product quality and access to customer evidence all affect what can be learned.
Choose one primary outcome before choosing channels. An online store might care about completed purchases from a specific product category. A service business might care about qualified consultation requests. Define what qualified means so the agency does not optimize toward cheap but irrelevant form submissions.
Days 1–30: establish the baseline and remove friction
Document the current offer, customer segment, inquiry process and available analytics. Test the main conversion action on a phone. Confirm that a submitted inquiry reaches someone who can respond and that duplicate events are not inflating reports. Record known gaps rather than presenting incomplete data as a precise baseline.
Interview the founder and, where permission allows, a few customers. Collect the language people use when explaining the problem, their objections and the alternatives they considered. Turn that material into one improved landing page and a short content brief. Agree on an approval owner so a campaign does not stall between departments.
Days 31–60: run a few comparable experiments
Test a limited set of messages against the same offer. For example, publish an explanation of the problem, a product walkthrough and a practical checklist. Label this as an experiment, not proof that a particular content format always wins. Track which visits and inquiries can reasonably be attributed to each piece.
Keep a simple experiment log: hypothesis, audience, asset, destination, dates, cost and result. Record qualitative feedback alongside counts. If visitors misunderstand the offer, rewrite it before multiplying distribution. If content attracts the right audience but inquiries go unanswered, fix response ownership before asking the agency for more traffic.
Days 61–90: decide what to repeat, change or stop
Review results against the original objective. Keep experiments with useful evidence, investigate uncertain ones and stop work that consistently attracts the wrong audience. Small samples should lead to cautious conclusions. A single sale does not establish a repeatable acquisition channel, and one quiet week does not prove a channel cannot work.
The next scope should name the work you will repeat, the person who maintains it and the questions still unanswered. For SEO and AI-search work, report improvements to pages and discovery separately from business outcomes. Google says established SEO practices apply to its AI Search features, but eligibility does not guarantee inclusion. A delivery calendar is something an agency can commit to; a search ranking is not.