Running Agency Operations on HubSpot Instead of Six Different Tools

Here is the step that costs an agency the most money, and it is not glamorous. A project manager writes the same status update three times: once in the project tool, once in the client's Slack channel, once as a note on the deal. Eleven clients, twice a week. That is most of a day, every week, produced by nobody deciding where the status actually lives.

Multiply that by the other copy and paste jobs, the sold scope living in a proposal PDF, the hours living in a timesheet tool, the renewal date living in someone's calendar, and you get an agency that cannot answer a simple question: which of our service lines makes money.

This is what HubSpot is for once you are past about sixty people, and Operations Hub is the part that does the connecting.

What Operations Hub actually does

Four things, and it helps to be specific because the marketing around it is vague.

• Data Sync. Two way sync between HubSpot and other tools, so a change in one is reflected in the other without anyone retyping it.
• Data quality automation. Format property values, standardise phone numbers and country fields, merge duplicate records.
• Programmable automation. Custom code inside a workflow, for the logic that does not fit if this then that.
• Workflow extensions. Workflow actions that reach into another system.

Those four are what turn a CRM into the place your delivery process runs, rather than a record of what sales sold.

Build the Closed Won trigger first

One workflow, and it pays for the whole exercise. When a deal in Sales Hub moves to Closed Won:

1. Create the project or board in your project management tool
2. Create and assign the delivery team's first tasks
3. Send the welcome email and the onboarding pack
4. Book the kickoff

The reason to start here is that this handover is where an agency loses client goodwill fastest, because the client has just paid and then hears nothing for eight days while somebody works out who owns the account.

Define your service lines so you can report on them

This is the unglamorous foundation, and skipping it is why most agencies cannot answer the profitability question.

Use custom properties and the product library to define what you sell as actual records: monthly SEO, website build, paid media management, content package. Then associate them with the deal. That gives you a structured, reportable answer to what was sold to whom, which then gives you:

• Forecasting by service line rather than one blended pipeline number
• Renewals and upsells you can actually list
• Reporting that survives the person who remembered it all leaving

Name them once and never freehand a variant. Two properties called Paid Media and Paid media management will split every report you build for the next three years.

Watch delivery against what was sold

Link tasks and Service Hub tickets to the deal or the company, and the reporting follows:

• Delivery progress against what the client bought, for both fixed scope projects and ongoing monthly work
• SLA and response times
• Where work is stalling, by stage rather than by feeling

The useful version of this is a list, not a chart. A list of accounts where delivery has stalled is something an account director can act on this week. A chart showing average delivery time across the agency is something everyone nods at.

Work out which service lines make money

Put the two halves together. From Sales Hub you have deal values, win rates and the service mix. From Service Hub and tasks you have hours logged, tickets resolved, CSAT and NPS. Together they answer three questions:

• Which services earn the most for the effort they take
• Which monthly agreements are underpriced or over-serviced
• Which acquisition channels bring clients who stay

Most agencies find at least one service line they have been running at a loss because it was easy to sell. The five dashboards worth having are covered in a separate piece on running revenue reporting across a growing team. If sales and delivery capacity are the part you are trying to join up, there is also a write up of how HubSpot and Forecast.app fit together.

Who owns the workflows you build for a client

This one catches agencies out, and it belongs in an operations piece because it is an operations decision, not a legal afterthought.

IP Australia's default position is plain: IP created by a contractor is the property of the contractor unless otherwise stated in the contract. Employers own the IP their employees create in relation to the business.

So the workflows, custom coded actions, report definitions, templates and integrations your team builds inside a client's portal belong to your agency by default. Not to the client. Unless your contract says otherwise, which most agency contracts do not, because nobody thought of the workflow as IP.

Two practical consequences.

First, decide which way you want it and write it into the contract either way. Assigning it to the client is a perfectly reasonable commercial choice. Keeping it and licensing it is also reasonable. Silence is the bad option, because silence means the client is operating on an assumption that is wrong, and they will find out at the worst possible moment.

Second, keep a record of what you built and where. When a client leaves, an agency that can hand over a documented list of workflows and custom properties keeps the relationship. One that cannot leaves the next agency to reverse engineer it, and that story gets told.

Connect the rest of the stack

No agency runs on one platform. Data Sync connects the specialist tools you keep:

• Accounting, such as QuickBooks
• Project management, such as Asana, Trello or Monday.com
• Communication, such as Slack or Teams

The two way part is what removes the retyping. A project status change in the project tool updates a property on the deal or ticket. A payment marked as paid in the finance system updates the client's status. That is the same status update from the top of this piece, written once.

Rolling it out without stopping client work

Fix the process first, then automate it. Automating a broken handover just makes it break faster.

Audit what you have. Map the current process across sales, delivery, service and renewals, then ask three questions: which manual tasks take the most time, where does data get duplicated or lost, and where do the delays actually happen. That gives you an ordered list rather than a wish list.

Do one or two processes, not everything. Client onboarding and data cleanup are the usual first pair. One service line or one team, prove it, then extend. Quick wins here are not a consolation prize, they are how you get the rest of the agency to cooperate.

Train people on the why. Explain what changes about their day, use HubSpot Academy for the mechanics, and pick someone in each team who is the person others ask. Adoption is the whole game. A perfectly built portal that people work around is worse than the spreadsheets, because now there are two versions.

Where to start

Pick the one report you cannot currently produce, usually profitability by service line, and work backwards to what has to exist for it to be possible. That gives you the properties, the product library entries and the associations, in order, with a reason attached to each.

If you are weighing up doing this with a partner, the tiers are explained in how HubSpot partner levels work in Australia, and this kind of build is most of our RevOps and HubSpot work. Or tell us which report you cannot produce and we will tell you what is missing.