Gemini Prompts for Marketing Plans

The short version: Gemini can build you a decent first-draft marketing plan if you feed it specific numbers, real constraints and a structure to follow, but it will happily produce confident nonsense if you ask vague questions. The prompts below are built around giving Gemini less room to guess and more reason to ask you questions back.

I use Gemini alongside other models most weeks for client strategy work, and the pattern I keep seeing is the same one that trips people up with every AI tool: they type "write me a marketing plan for my business" and then wonder why they got back something that could apply to literally any business. A marketing plan is only useful if it's specific to your numbers, your customer, your budget and your timeline. Gemini doesn't know any of that unless you tell it, so the quality of what you get out is almost entirely down to the quality of what you put in.

Why most marketing plan prompts fail before you even hit send

The biggest mistake is asking Gemini to do everything in one prompt. "Create a full marketing plan including strategy, channels, budget, content calendar and KPIs" sounds efficient but it forces the model to spread thin across six jobs at once, none of them done well. You end up with a document that looks thorough but has no depth anywhere.

The second mistake, and this is the uncomfortable bit nobody likes saying out loud, is that a marketing plan written entirely by AI with no input from someone who knows the business is worth roughly nothing. Gemini doesn't know that your best customers come from referrals, not ads. It doesn't know that your team can't manage a daily posting schedule because there are two of you. It will invent assumptions to fill those gaps, and if you don't correct them, you'll build a plan around a business that doesn't exist. Treat Gemini as a drafting partner that needs correcting, not an oracle.

Start with context, not the ask

Before any prompt that asks for output, give Gemini a block of context it can refer back to. This is the single biggest lever for quality and almost nobody does it.

• What the business sells and to whom, in plain terms
• Monthly marketing budget, even a rough range
• Current channels and how each one is performing
• The main business goal for the next 90 days (revenue, leads, sign-ups, whatever it is)
• Team size and who will execute the plan

If you're running this inside a longer project, this context block is also what makes Gemini's scheduled actions worth setting up, because you can have it check performance against the plan weekly without re-explaining the business every time.

The worked example: a made-up small business

Say you run a small online shop selling handmade candles, turning over about £8,000 a month, with a marketing budget of £500 and no in-house content person. Here's how the prompting would go, step by step.

Step 1, the context prompt:

"I run an online candle business. Average order value is £28, monthly revenue is around £8,000, monthly marketing budget is £500. Current channels: Instagram (4,200 followers, low engagement), email list of 900 people, no paid ads yet. Goal for the next 90 days is to grow revenue to £11,000 a month. It's just me running marketing, maybe 6 hours a week available. Remember this for the rest of our conversation."

Step 2, the diagnosis prompt:

"Based on what I've told you, what are the three most likely reasons revenue has plateaued, and what would you need to know from me to confirm them?"

This is the step most people skip. Gemini will come back with things like "is the Instagram engagement drop due to a change in posting frequency, or an algorithm shift, or a change in content type?" and will ask whether the email list is being emailed regularly. Answer. This is where the plan starts being about your business rather than a template.

Step 3, the plan prompt:

"Using everything above, build a 90-day marketing plan with three channel priorities only, not six. For each priority give me: the specific weekly action, the time it takes, and one measurable target. Budget split should total £500 a month. Assume 6 hours a week of my time."

For this example, a realistic Gemini output might prioritise email (because the list already exists and converts better than cold Instagram traffic), a small £300 a month Meta ads test targeting past purchasers and lookalikes, and a simplified Instagram approach of two posts a week instead of daily, freeing time for the other two. That's a usable plan because it respects the real constraint, 6 hours a week, instead of pretending you have a marketing department.

Step 4, the stress test prompt:

"What's the weakest assumption in this plan, and what would make it fail?"

Always ask this. It's the single highest-value prompt in the whole sequence because it forces Gemini to critique its own output instead of just defending it, and it usually surfaces something like "this assumes the email list is still engaged, but you haven't emailed them in a while, so open rates may be lower than typical."

Prompts for each section of a plan

If you'd rather build section by section than in one flowing conversation, these work well on their own, always after you've set the context block:

• Positioning: "Write three different positioning statements for this business, each targeting a different customer motivation (price, quality, convenience). Keep each under 25 words."
• Channel selection: "Given my budget and time, rank these five channels by likely return in the next 90 days: email, organic social, paid social, SEO content, partnerships. Explain the reasoning for the top two."
• Content calendar: "Build a 4-week content calendar for [channel] with one post per [day], each with a one-line hook idea, not the full copy yet."
• Budget allocation: "Split a £[amount] monthly budget across the channels we've agreed, with a one-sentence justification for each split."
• KPIs: "Give me 3 KPIs for this plan that I can measure with tools I already have, not vanity metrics."

Keep each of these as a separate exchange rather than cramming them into one mega-prompt. Gemini's answers get noticeably sharper when it's solving one problem at a time rather than juggling five.

Where Gemini helps and where it doesn't

Gemini is good at structure, at generating options you hadn't thought of, and at pulling a messy brain-dump of facts about your business into something organised. It's also decent at playing devil's advocate if you explicitly ask it to, which most people never do.

It is not good at knowing your market the way you do. It doesn't know that your competitor just dropped their prices last week, or that a key supplier is about to raise costs, or that your best-performing ad last quarter was a fluke tied to a seasonal event. These are the things that change a marketing plan, and you have to feed them in manually every time, because Gemini has no live awareness of your specific market unless you tell it. This is also true of competitor tools, and it's worth knowing that even comparing something like DeepSeek vs Gemini on data handling matters if you're pasting real revenue and customer numbers into either one, because the context block above only works if you're comfortable with where that data goes.

If you're deciding which Gemini tier is worth paying for to do this kind of work regularly, it's worth reading up on what replaced the old Gemini Advanced plan before you commit, since the naming and features have shifted and the free tier has real limits on longer, multi-step planning conversations like the one above.

A short word on research inputs

A marketing plan is only as good as the research feeding it. If you're pulling together customer interviews, competitor notes or long reports before you prompt Gemini, a tool like NotebookLM, now folded into Gemini's ecosystem, can help you summarise that material first. It's worth knowing what the free plan on Gemini Notebook limits before you rely on it for a big research pile, because the caps on sources and length catch people out mid-project.

When to stop using Gemini and bring in a person

If your plan involves a budget over roughly £5,000 a month, multiple stakeholders who need to agree on direction, or a market you don't understand well yourself, that's the point where a human strategist or an AI consultant for small businesses earns their fee. Gemini can draft, structure and stress-test, but it can't sit in a room, read the politics of a decision, or take responsibility when the plan doesn't work. That accountability is the part you're paying a person for, not the writing.

A quick checklist before you trust any Gemini-built plan

• Does every number in it match a real figure you gave it, not an assumed one?
• Have you asked it to name the weakest assumption?
• Does the time required match the actual hours available, not an idealised version of your week?
• Would you be comfortable explaining every line of this plan to a colleague without saying "the AI suggested it"?

If the answer to that last one is no, it's not ready. Go back and push on whichever section feels thin.

Frequently asked questions

What is the best Gemini prompt for a marketing plan?

There isn't one single best prompt. The approach that works is giving Gemini full context about your budget, team, goals and current channels first, then asking it to diagnose problems before it writes any plan, then stress-testing the output by asking what assumption is weakest.

Can Gemini write a full marketing strategy on its own?

It can draft one, but it will fill any gaps in information with assumptions that may not match your real business, so a plan built without your corrections and specific numbers should be treated as a first draft, not a finished strategy.

Is Gemini better than ChatGPT for marketing plans?

Both can produce solid drafts if you prompt them with specifics rather than vague requests; the real difference tends to be in how each handles longer context and ongoing conversations, so it's worth testing both with the same business details and comparing the outputs directly.

How detailed should my prompt be for a marketing plan?

Very. Include your budget, revenue, team size, available hours per week, current channels and a specific 90-day goal, since each of these changes what a sensible plan looks like and Gemini has no way to guess them accurately on its own.