Build Your Marketing Plan From the Customer Data You Already Have in HubSpot

Your marketing plan for next year is already written. It is sitting in your HubSpot portal as 18 months of closed deals, and nobody has read it.

Every won deal in there recorded where the contact came from, how long it took and what it was worth. Every lost one recorded why you lost it, or at least it did if somebody filled the field in. That is a better place to start than any trends article, this one included.

What your own HubSpot data already tells you

Pull 18 months of closed deals and break them down by source

Build a custom report across Deals and Contacts. Filter on Close date in the last 18 months, separate closed won from closed lost, and group the won ones by the Original source of the contact attached to the deal.

Three numbers come out of that, and they are the three that decide a budget:

• How many won deals each source produced
• The average Amount of a won deal from each source
• Days to close, by source

A source that produces plenty of small deals quickly is a completely different budget decision from one that produces four deals a year at ten times the size. Most plans treat them as the same line item and then wonder why the forecast is wrong.

If your portal has attribution reporting, use it. If it does not, the Original source and Original source drill-down properties on the contact record will get you most of the way. The CRM data trends actually worth planning around are the ones you can see in your own numbers first.

Read your closed lost reasons before you write a single campaign brief

Closed lost reason is a default deal property, and in most portals we open it is empty. Fixing that takes an afternoon: in pipeline settings, add it to the closed lost stage under Conditional stage properties and mark it required, so a rep is asked for it on the way out.

Once you have three months of honest answers, group them. Price. Timing. Went with a competitor. No budget this year. Went quiet.

Each group is a different piece of marketing work. Price objections are a proof problem, so you need customer numbers you are allowed to publish. Timing is a nurture problem. Losing to a named competitor is a positioning problem, and using competitor data alongside your own customer data tells you which one keeps beating you.

This is also the cheapest research you will ever run. You are already having the conversations. You are just not writing down how they ended.

Read the last hundred support tickets

Closed lost tells you why people did not buy. Tickets tell you what happens to the ones who did, and the same three complaints will be sitting in there in different words.

Filter tickets by category over the last 12 months and count them. If the biggest category is people asking how to do something, you have an onboarding and documentation job that marketing can help with, and you have the exact words customers use for it. Those are the words that belong on the landing page.

Same goes for survey responses if you run NPS or a post purchase survey in HubSpot. A low score with a written comment attached is the most useful paragraph in your portal, and almost nobody in marketing ever reads one.

Describe your best customers as properties, then save it as a list

Filter contacts on Lifecycle stage is Customer, then look at the companies behind them: industry, employee count, state, and the amount they closed at. Take the top quarter by amount and write down what they have in common.

Save it as an active list. That list is your paid targeting, your email segment and the seed for any ad audience you build, and it updates itself as you win more. A persona written on a whiteboard two years ago does none of that.

Find the stage where deals stop

Run the deal funnel report over the same 18 months. One stage will be holding more deals than the rest and moving them slower than the rest. That stage is your content plan.

If deals stall between first meeting and proposal, what you are missing is proof and pricing clarity, and another six top of funnel blog posts will do nothing about it.

Using AI on that data without getting yourself in trouble

Here is the real number, and it is smaller than the noise around it suggests. The ABS found that 12 per cent of Australian businesses used AI in 2024 to 25. Among medium sized businesses it was 22 per cent, and among large businesses 35 per cent. Three years earlier, before ChatGPT existed, the figures were 3 per cent of medium businesses and 9 per cent of large ones.

So if you are pointing AI at your customer data this year, you are early. Early is fine. Early with no rule about what goes into the tool is how a customer list ends up in a chat window that nobody owns.

There is a current Australian framework for this and it is short. The National AI Centre published its Guidance for AI Adoption on 21 October 2025, built around six essential practices: decide who is accountable, understand impacts and plan accordingly, measure and manage risks, share essential information, test and monitor, and maintain human control.

Two of the six matter in the first week. Maintain human control, so somebody reads AI written segment logic and campaign copy before it sends to your whole database. And test and monitor, so you check an AI built list against a manual count before you spend money against it.

The other four are worth an hour of somebody's time before you connect a new AI tool to the CRM, because that is the moment the decision gets made and nobody revisits it.

Turn it into a plan with numbers on it

Now you can write the plan, and every line has a figure from your own portal behind it:

• Customer acquisition cost, or CAC, by source, from the spend you can attribute and the deals you counted
• Average deal amount by source, so a channel target is in dollars
• Stage to stage conversion, so a lead target is a real number rather than a wish
• Days to close, so you know which quarter this spend shows up in

Group each campaign's assets in HubSpot's campaign tool so the reporting comes back grouped the way you planned it. Then build the reporting once. The five dashboards that turn this data into a plan covers what belongs on them, and keeping a campaign cohesive covers the part where eight assets all have to say the same thing.

Nailing the basics first

None of this works on a portal where the data is wrong, and the fixes are boring:

• Closed lost reason attached to the stage, and filled in
• Original source protected from imports and integrations, because an import that stamps everything as Offline sources erases where people came from
• One deal per opportunity, one owner, and stages that mean the same thing to every rep
• Close date used as an actual close date, so a reporting period means something

Do those before the planning session, or you will spend the session arguing about whether the numbers are real. The CRM fixes to make before you build next year's plan goes through the rest of them.

If you are still choosing the system underneath all of this, our HubSpot and Salesforce comparison is the honest version.

The question worth asking in the planning session

Open the deals report, find the source behind your three biggest won deals last year, and ask what share of next year's budget it is getting.

If those two percentages are nowhere near each other, you have found your plan. Give us a tap if you want a second pair of eyes on the report.