Best Practices & Playbooks

The operating rhythm, four complete campaign playbooks you can copy, how agencies run multiple brands, and the mistakes that cost the most.

Jump to module

Lesson 1: The Operating Rhythm

A campaign works best on a weekly rhythm with a small number of fixed appointments. The rhythm matters more than any setting on the page.

Setup (once, about fifteen minutes)

  1. Pick the angle. Take the recommendation unless you have a reason.
  2. Read the theme before the name. The theme is what the content will actually be about; a wrong theme is a wrong campaign no matter how good the title.
  3. Write context notes if the campaign is about something specific — the launch, the real deadline, the number you want led with.
  4. Attach a source if there is a page that describes the thing.
  5. Set the accounts. Check each one is genuinely connected.
  6. Set the duration from the subject.
  7. Set the mix last, from your review capacity — then multiply by the weeks and look at the total.

Weekly (about twenty minutes, one sitting)

  • Decide on the week's drafts. All of them. Decide, do not browse — an undecided draft is the same as a rejected one and cost you a generation.
  • Skim the timeline to check the content still matches the angle.
  • Read the last error if one is set, before doing anything else.

Weekly (five minutes, the calendar)

  • Scan the week ahead for collisions, gaps and overload across all your content, not just this campaign's.

At the end (about ten minutes)

  • Read the outcome line and the empty reason if there is one.
  • Decide whether the angle earned another run. That is the question a campaign exists to answer.
  • Duplicate if it did; pick a different angle if it did not.

The one habit that matters most

Put the weekly review in your diary as an appointment. "Awaiting approval" is the most common reason a campaign ends having published nothing — the content existed, it was fine, and nobody decided. Twenty minutes a week fixes more campaign problems than every setting in this course combined.

Lesson 2: Four Playbooks

Copy these directly. Each names the angle, the shape and the specific thing people get wrong.

Playbook 1 — The Product Launch

Angle: Problem / Solution Duration: 3 weeks Mix: 3 image posts, 1 carousel per week Total: 12 pieces

Setup: Context notes carry what is launching and what it replaces. Attach the launch page as a source. CTA points at the page, not the homepage.

The shape: Week one names the problem without mentioning the product. Week two shows the fix. Week three is proof — what changed for whom.

How to steer it: Put the arc in the context notes explicitly. "Week one: the problem, no product mention. Week two: how we fixed it. Week three: results." The notes reach every week, so this is the one field that can carry a three-week arc.

Most common mistake: Launching with Offer & Promotion. That angle leads with the deal and assumes people already understand the thing. If nobody knows what it is yet, you are promoting to an audience that has no context.

Playbook 2 — The Promotion With a Deadline

Angle: Offer & Promotion Duration: 1–2 weeks Mix: 3 image posts, 1 text post per week Total: 4–8 pieces

Setup: The real deadline goes in the context notes. Bolta will not invent urgency, but it will use yours. The CTA is the booking or purchase link.

The shape: Short and dense. A promotion that runs four weeks is not a promotion.

How to steer it: State exactly what the offer is, what it costs, and when it ends. Vague offers produce vague posts — this angle leads with the offer itself, so if the offer is thin, so is every post.

Most common mistake: Running it too long. The window should match the offer window. If the deal ends on the 30th, so does the campaign.

Playbook 3 — Thought Leadership Without Overclaiming

Angle: Thought Leadership Duration: 4 weeks Mix: 2 image posts, 1 carousel per week Total: 12 pieces

Setup: This angle needs material more than any other. Business DNA filled in, a source attached, and context notes carrying actual things you have done or seen — a decision you made, a result you got, a thing you changed your mind about.

The shape: Slower cadence, higher density per piece. Twelve strong posts beat twenty-four safe ones.

How to steer it: The angle demands a position a reasonable person could disagree with. If your context notes contain no opinions, you will get confident generalities. Write the positions down; let Bolta develop them.

Most common mistake: Running it on an empty workspace. With no material, this is the angle that produces the most plausible-sounding content that says nothing.

Playbook 4 — Consistency Test for a New Business

Angle: Quick Tips or Product Education Duration: 2 weeks Mix: 3 image posts per week Total: 6 pieces

Setup: Minimal. This is deliberately a small campaign — it exists to find out whether you can sustain a review rhythm and whether the angle lands.

The shape: Two weeks, modest mix, one platform if you are unsure.

How to steer it: Do not over-configure. The point is to establish the loop — generate, review, publish, read the result — before you commit to anything bigger.

Most common mistake: Starting with a four-week, five-format, three-platform campaign. Twenty-plus drafts on week one of using the feature is how people conclude it does not work for them.

Choosing between them

If you haveStart with
Something specific launchingPlaybook 1
A real deadline and an offerPlaybook 2
Opinions and a track recordPlaybook 3
None of the above, or you are newPlaybook 4

Lesson 3: Multi-Brand and Agency Work

Campaigns are built to run several brands from one place, and two features do most of the work.

Per-campaign voice. Each campaign names the voice it writes in, carried to every producer that week. One workspace can run a founder-voice campaign and a brand-voice campaign in the same week without swapping a global setting between them.

Per-campaign settings. Sources scope, media scope, tone and image quality can be set for one campaign only. Restrict a client's campaign to that client's own media and sources so nothing leaks between accounts.

The agency pattern

  1. Build the first client campaign carefully — angle, plan, settings, voice.
  2. Duplicate it for the next client.
  3. Change three things: the voice, the accounts, and the context notes.
  4. Leave the plan and settings alone. The shape is the part worth reusing.

Reporting

Campaigns are an analytics dimension as well as a producer. You can group and filter by campaign using the same derivation the campaign page uses — one definition of "published", everywhere.

Two things to say out loud in a client report:

  • Published is counted per account. A partial publish is reported honestly rather than as zero.
  • Per-campaign numbers do not sum to the total. A post can belong to two campaigns, so the workspace figure is derived over the union, not the sum. Presenting the sum as a total is wrong, and Bolta will not do it for you.

Attaching work Bolta did not make

You can add existing posts to a campaign. That is how content you wrote by hand, or published before the campaign existed, gets counted in the result. For client reporting this matters — the campaign should reflect the whole effort, not only the automated half.

Lesson 4: The Expensive Mistakes

1. Reading the mix as a total. By far the most common. "3 images and a carousel" over four weeks is sixteen pieces, not four. People plan a busy mix for a long window, get four times the drafts they expected, review week one and abandon the rest. Multiply before you commit, and size the mix from what you will actually read.

2. Treating a campaign as a schedule. A campaign is a push with an ending. If you want "keep posting three times a week forever", that is a Routine. Using campaigns for it means recreating one every month and never getting a readable result from any of them.

3. Not checking the accounts. A campaign armed with no valid account generates nothing and publishes nothing. It tells you — the phase says "needs accounts" — but it is a wasted week if you do not look. Connections expire quietly. If a campaign has gone silent, check connections first.

4. Letting the queue be the bottleneck. Generation is rarely the problem. Approval is. See the rhythm in Lesson 1.

And four smaller ones

Planning a format you cannot produce. Emails are written but never sent; blogs need a connected site; stories and avatar videos are not made at all.

Editing the plan mid-campaign and expecting it to apply backwards. Changes take effect from the next week that has not been laid out. Week one is already written.

Duplicating a campaign that failed on its angle. Duplicate carries the angle over. If the angle was the problem, start fresh.

Running two campaigns on the same subject. They will produce two uncoordinated versions of the same message. Let one campaign own a subject for its window.

What to take away

  • Two appointments: the setup, and twenty minutes a week in the queue.
  • Four playbooks to copy; pick by what you actually have.
  • Agencies: build once, duplicate, change voice + accounts + notes.
  • The four expensive mistakes are the mix, the schedule confusion, the accounts, and the queue. Three of the four are visible on the page.

Where to go next

  • Voice DNA & Brand Agents — the voice a campaign writes in, and how to train it.
  • Mastering Bolta Agents — the workforce that runs alongside your campaigns.
  • Cross-Platform Repurposing with Bolta — turning one campaign's content into more surface area.