How to use Gojiberry for a focused first campaign

1. Prepare a brief before configuring the product

Before deciding how to use Gojiberry, write down one buyer role, one account segment and one problem your offer solves. Include a short description of what makes a company unsuitable. This brief gives you a standard for judging results later. Without it, each plausible profile can tempt you to broaden the campaign until the audience no longer has a coherent reason to respond.

An illustrative brief might target an operations leader at a growing B2B service company that needs a particular process improvement. Replace that example with your actual market and evidence. Include a few existing good-fit accounts, but do not assume that every characteristic of those customers is causally important.

The steps below are recommendations based on official help documentation checked November 9, 2026. Named navigation paths come from those documents. We have not accessed a live Gojiberry workspace, and screens or labels may change. Use the current help article or in-app support if your interface differs.

2. Give the AI accurate company and offer context

Complete your workspace information before asking the system to generate messages. The AI messaging guide identifies company context as an input. Describe the product in language your buyer would use, the specific operational problem, the outcome you can support and the next step you want a prospect to consider.

Separate proof from aspiration. If you have a case study, state what it actually demonstrated and avoid presenting one customer’s result as a universal expectation. If you do not have evidence for a quantitative claim, leave the number out. Add tone instructions that a reviewer can check, such as concise, direct and limited to one question.

Create a short internal list of prohibited assumptions. For example, the AI should not claim a prospect is replacing a tool, struggling with a problem or actively purchasing unless you have evidence. This list becomes a useful review checklist even if you express the instruction through a template rather than a dedicated field.

3. Configure one audience and a small number of sources

Use the ICP definition to specify relevant roles, company characteristics, locations and exclusions. Start with the criteria that directly determine whether you can sell successfully. Avoid adding many mandatory phrases simply because they sound relevant; profiles may describe the same responsibility in different language.

Choose a source that expresses your commercial hypothesis. A topic-related signal is useful only if interest in that topic meaningfully relates to your offer. A lookalike source is useful only if the reference represents the type of customer you can repeatedly serve. Begin with a small setup so you can attribute the results to a specific choice.

The lookalike help guide documents a path through Agents, the selected campaign, Sources and Add leads source, followed by Warm Lookalike. Treat this as documented navigation rather than a guarantee of current labels. For any source, record the reference or signal and why you chose it before the first batch arrives.

If the system has an option to broaden discovery, decide whether that fits the experiment. A precision test is easier to interpret when you know which leads came from the intended source. Broader discovery may be useful later, but it should be an explicit choice.

4. Check senders before activating outreach

Before connecting LinkedIn, review its official policy prohibiting third-party automation of activity on its website. Scheduling, sending limits and human approval do not establish platform authorization. Decide whether that conflict is acceptable with the person responsible for the account; consider other appropriate channels if it is not.

Confirm the identity of the LinkedIn account or mailbox that will send and the person who owns replies. Check connection status, active days and the configured time window. Sender permissions and limits are operating constraints, not settings to maximize automatically.

For email, use your current mailbox and domain configuration as the starting point for the vendor’s deliverability checklist. The official mailbox guide includes authentication, reply delivery, sender identity and unsubscribe controls. Have the person responsible for your domain review any changes; a copied example record may be inappropriate for a domain that already supports other sending services.

Keep initial volume conservative and observe the resulting queue. Connecting a mailbox does not prove that messages land in the inbox, and a platform health indicator does not replace actual recipient-side evidence. If you need help interpreting a failed check, resolve it before using prospect responses as a test of your offer.

If you only need email, the help center documents an Email-Only Agent. Choose a channel configuration that matches the campaign instead of adding LinkedIn steps merely because they are available. Every extra step should have a reason from the buyer’s perspective.

5. Enable review and inspect the first leads

The Review Mode help article documents opening an agent’s Settings, enabling Review Mode and saving. Pending leads can then be reviewed from the Review tab. The important behavior is that prospects wait for approval before entering outreach. Check the current article if you cannot find the setting.

For each initial lead, inspect the role, company, source evidence and message. Use three simple outcomes in your own notes: accept, reject for fit, or research further. Record the reason in plain language. A pattern of wrong company sizes requires a different correction from a pattern of irrelevant signal interpretation.

Approve only after understanding what approval releases. If it activates the rest of a sequence automatically, review the sequence as a whole. Do not bulk-approve merely to clear the queue before you have checked a representative sample. A clean approval queue is less important than a consistent decision standard.

The first-lead help guide notes that targeting updates take time to affect future batches. After a change, distinguish newly discovered leads from the previous queue. Otherwise you may conclude that an adjustment failed by reviewing records created under the old settings.

6. Test messages and sequence logic

Preview generated messages across several accepted profiles. Check facts first, relevance second and wording third. A message can be fluent yet inappropriate if it assumes a need the prospect never expressed. Remove unsupported personal details and ask a useful question rather than over-explaining your company.

A recommended opening structure is a brief context line, a clear business reason to speak and one easy-to-answer question. This is editorial guidance, not a vendor-prescribed template. Avoid turning every observed event into an explicit statement that you monitored the person.

Read every follow-up in order. Check whether it makes sense if the prospect has not accepted a connection, if email is unavailable or if the person replies early. Confirm the current stopping behavior in your workspace. Assign a human to review conversations and progress useful responses so automation does not end at an unattended inbox.

For a controlled evaluation, keep the offer stable while testing one meaningful difference. If you change the audience, source, message and timing together, you will have little basis for deciding what to keep.

7. Diagnose a quiet campaign in the right order

A quiet campaign can indicate a sourcing issue, an approval hold, a sender problem or a normal workflow dependency. Check which stage is actually empty before changing the copy. The official troubleshooting guide separates active campaign status, connected senders, available leads, shared limits and Review Mode.

Do not disable review just to make activity appear. If prospects are waiting for your decision, inspect and approve the suitable ones. If there are no candidates, inspect the source and filters. If there are accepted leads but no available action, check the schedule, sender capacity and dependencies such as invitation acceptance.

The help center’s older first-reply article and newer scheduling articles differ on launch frequency. We use the newer October 9 scheduling guide for the description of twice-daily outreach, but your configured account schedule is the operational reference. Avoid interpreting a short wait as proof that the system is broken.

8. Review results before expanding

Summarize the initial campaign with a few meaningful measures: suitable prospects reviewed, proportion accepted, time spent reviewing, substantive conversations and agreed next steps. Keep stop requests and negative feedback visible. These observations explain whether the workflow deserves more volume.

Expand the component that is genuinely constrained. Better source coverage may help when the existing audience is relevant but too small. Better messaging may help when suitable buyers do not understand the offer. Additional senders may matter when approved leads wait behind an appropriate capacity limit. Each diagnosis implies a different purchase or process change.

Before adding recurring spend, revisit billing terms and the renewal date. The end of a trial is a decision point, not evidence that the workflow has succeeded. Keep the configuration only if you can explain the value it is producing and who will continue operating it.

Frequently asked questions