Is X worth it for agencies and consultants? When it makes sense and when it doesn't
Is X (Twitter) worth it for agencies and consulting firms: how to check whether your buyers are there, X for research or sales, a 90-day test and when to quit.
Should an agency or a consulting firm be on X (formerly Twitter)? First, check whether your buyers post there: list 20 people from companies in your ICP and see how many of them posted on X in the last 30 days about problems you solve. If fewer than three, use X for research: read discussions in your clients’ industry and note their questions, but don’t count on sales. If three or more, run a test: for 90 days, post short observations and comments on news from your clients’ industry, reply to buyers’ posts and count conversations with people from your ICP. If after 90 days you have none, pause the channel and write down why.
In this post I’ll show how to check whether X makes sense for your firm, how to use it in both roles and when to quit.
In US data, X is far behind LinkedIn
In Pew Research Center data from 2026 (a survey of 5,511 US adults), 22% of US adults with a college degree use X, compared with 51% who use LinkedIn. In the Content Marketing Institute’s research for 2025 (980 B2B marketers, mostly from North America), 7% of respondents named Twitter as the platform that delivers the best value for their organization, compared with 85% for LinkedIn.
This is data from the US and North America, not from every market; I didn’t find current data on who uses X in Poland. In my view, though, it shows that X isn’t a default channel for a service firm, but a channel to check in your specific niche.
Check whether your buyers post on X
List 20 people who decide on buying your services at companies in your ICP: directors, owners, department heads. Search for them on X by first name, last name and company name, and if you can’t find them, check whether their LinkedIn profile or company website links to an X account. For each person, note whether they have an account, when they last posted and what they write about.
I suggest a simple threshold: if at least three of the 20 wrote about problems you solve in the last 30 days, X is worth a test. If fewer, X won’t be your sales channel. People from your own industry (other agencies and tool vendors) don’t count toward the result.
X as a research tool
Even when few buyers are on X, it can be useful as a source of knowledge. Create a list on X with people from your clients’ industry, industry journalists and analysts. Once a week, spend 15 minutes reading the posts from that list and note the questions, complaints and topics that keep coming up. Those notes become topics for blog posts and LinkedIn posts.
In this role, you don’t have to post anything, and the result can be better topics for the channels where your buyers are. If you do run the test, the same list will show you which buyers’ posts to reply to.
X as a sales channel: short posts and replies
If your buyers are on X, post in three forms:
- a short observation from a project or from your client’s market: one idea and one specific detail,
- a comment on news from your client’s industry: what it means for the companies you sell to,
- replies to buyers’ posts that add information instead of advertising.
Keep to a simple format: one idea per post, 280 characters at most, and educational material as a thread of 8–12 posts. I expect that in a service firm, replies to other people’s posts will lead to the most conversations, because they reach someone who is writing about their problem right now; the test will tell.
The 90-day test: four decisions before you start
Before you start, write down four things:
- Cadence and time: for example, two posts and five replies a week, about two hours.
- Who does it.
- The end date: 90 days from the start.
- The criterion: at least two conversations with people from your ICP mean the test worked; one conversation is a signal to extend the test by another 90 days; zero means pausing the channel. After an extension, you judge the same way, but one conversation now means pausing.
A conversation is an exchange of messages or a call about that person’s problem (when a discussion in the replies develops, ask whether you can message them), not just a like or a single reply to a post. Every week, record in a spreadsheet similar to the one in the post on LinkedIn for founder-led service firms, but with one row per week: how many posts and replies you published, how many people from your ICP replied or sent you a message, and how many of those contacts led to a conversation. You check whether someone fits your ICP by their company and job title.
Record impressions and followers if you have access to them, but don’t judge the test by them. Don’t end the test early: expect the first weeks to be mostly about building the list of accounts and your first replies.
When to quit and how to record it
If after 90 days you have no conversation with anyone from your ICP, you pause the channel: you stop posting, and the account stays if you use it for research. The hours from the test go back to the channels that work. If the test worked, the channel becomes active, with the cadence from the test.
Write the decision down so nobody starts from scratch six months from now. Keep a short file for each channel: its status (active, paused or closed) and a description of what you post there, how and how often. Paused means you’ll check again in six months; closed means you’re not coming back. A file with just a status and no description is only a bookmark.
Example: an agency’s decision
This is an illustration, not a real firm: an agency that runs campaigns for law firms. The owner listed 20 managing partners from law firms in their ICP. In the last 30 days, two of them had posted on X, both about court cases, not about law firm marketing.
| Step | Result (illustration) |
|---|---|
| Buyers on X | 2 of 20 post, and none writes about problems the agency solves: below the threshold |
| Decision | no test; X for research only: a list with legal journalists and law firms |
| Time | 15 minutes a week to read the list |
| Record | X channel: paused as a publishing channel, used for research; reason and check date in the channel file |
| Next | LinkedIn posts on topics from the list; check the 20 people again in six months |
Frequently asked questions
The company account or the owner’s?
I think that in a service firm, as on LinkedIn, a personal account works better, because replies to buyers’ posts are written by people. I described how to write for buyers rather than for your industry in the post on LinkedIn for founder-led service firms.
Should I run X and LinkedIn at the same time?
If you have time for one channel, choose the one where your buyers are. You can rework the same text into a short post on X, but you write replies to buyers’ posts separately.
Do I need a paid account?
Not for the test, in my view. A paid account won’t change whether your buyers are on X.
Summary
If you’re considering X, start with a list of 20 people from companies in your ICP and a threshold of three buyers who wrote about problems you solve in the last 30 days. Below the threshold, use X for research for 15 minutes a week. Above it, run a 90-day test with a written cadence and criterion, and after the test record your decision in the channel file. In our open repo (files and instructions for an AI agent such as Claude Code), every channel has a file with its status and a description of what gets published there, and the agent only proposes content that one of the active channels accepts (open source).
Every method on this blog also runs as a skill in the open-source B2BForce repo. About the repo