ChatGPT + HubSpot: compare two campaigns, then propose the next one

When to use it

You ran two campaigns, say a webinar and a downloadable guide, and need to decide which one to learn from before planning the next. The first prompt is built for ChatGPT's deep research mode with HubSpot enabled as a source, because a campaign comparison is a broader investigation you want cited. Campaign reporting depends on your HubSpot subscription; the prompt tells the model what to do if it is unavailable.

Prompt 1: compare the campaigns (deep research)

Select deep research and enable HubSpot as a source. The prompt insists on keeping sessions, new contacts, influenced contacts, associated deals and attributed revenue separate, because they mean different things. An influenced contact engaged with a campaign; an associated deal is connected to it; attributed revenue depends on the reporting model. You want those distinctions visible before a budget decision.

Use the HubSpot connector to compare the campaigns [CAMPAIGN A] and [CAMPAIGN B] for [DATE RANGE]. For each campaign show, where the data is available: sessions, new contacts, influenced contacts, associated deals by stage, and revenue or revenue attribution. Keep these metrics in separate columns and do not combine them into a single score. For each report, explain the attribution model and the date basis used. Identify contacts or deals that overlap between the two campaigns. Cite the HubSpot sources for every number and state clearly which data was missing or not accessible. Finish with one recommended follow-up experiment, separating what the evidence shows from what is a hypothesis. If campaign reporting is not available in this account, say so and instead compare accessible closed-won and closed-lost deals by their original source property, labelled as a deal-source comparison. Do not create or change anything in HubSpot.

Prompt 2: propose the next campaign from the findings

Now ask a content question with a reason behind it. If the evidence suggests prospects need help understanding implementation, the next offer might be an implementation workshop. That is a hypothesis you can test, so the prompt asks for a measurable success criterion.

Based only on the verified findings above, propose one follow-up campaign for [AUDIENCE]. Include: the customer problem it addresses, the offer, a three-email outline, the landing-page message, and one measurable success criterion with a definition of what counts as a qualified enquiry. For each recommendation, name the finding that supports it. Write the first email in full. Keep this as a proposal in chat; do not create drafts in HubSpot.

How to use it

1. 1Open a new conversation, select deep research, and enable HubSpot as a source.
2. 2Run prompt 1 with two real campaign names and a date range. Open one cited report in HubSpot with the same dates and settings to verify a number.
3. 3Point to one verified difference between the campaigns before moving on.
4. 4Run prompt 2 in the same conversation. Review the proposal, then carry the message into your email drafts or a landing page.

Tips

• Keep metric types separate. Mixing influenced contacts with attributed revenue is the fastest way to a wrong budget decision.
• Ask for the attribution model and date basis every time; two reports with the same name can use different ones.
• If email campaign actions are available in your account, the reviewed content can go straight into a HubSpot draft. Review before sending.
• Define what a qualified enquiry means before you launch, so the success criterion is measurable.

Related resources