What Should Marketing Do Next?

Key takeaways

• Start from the stage that leaks, not from the channel everyone is talking about.
• Judge a channel on three things: are your buyers there, can you afford it, and how fast will it tell you something.
• Test a few channels cheaply, then put most of your effort into the one that works.
• A fair run depends on the channel: weeks for paid ads and outbound, months for content and SEO, at least one sales cycle for anything aimed at revenue.
• Competitor research helps at two moments: before you pick a channel, and when a working channel stops working.

View full size

The planning meeting starts with a list. Someone wants to try TikTok, someone read that webinars are back, the CEO saw a competitor's billboard at the airport, and the agency has a proposal for a podcast. Every idea sounds reasonable, which is the problem; a list of reasonable ideas doesn't tell you which one to do first. What tells you is a short sequence of questions, asked in the same order every time.

Step 1: Start from the leak, not the channel

Before you choose any channel, look at your funnel stage by stage and find the one that is furthest below its usual level, or the step between two stages where the most people fall out. That's the leak. A new channel that fills the top of the funnel won't help much if the real problem is that qualified leads never book a demo.

Compare each stage with its own usual range over the last 8 to 12 weeks rather than with last week; the campaign guide explains why. Pick one leak at a time. Two leaks fixed at once usually means neither gets a fair test.

Step 2: Choose channels with three questions

Are your buyers already there? B2B buyers increasingly research on their own. In a 2025 Gartner survey, 61% of B2B buyers said they prefer a buying experience without a sales rep (Gartner), and an earlier Gartner survey found buyers value third-party sources such as peers and review sites more than suppliers' own digital channels (Gartner). Channels where buyers learn from others, like communities, partners, analysts and review sites, often deserve more weight than they get.

Can you afford it, in money and in time? Paid channels cost money from day one; content and community cost time and pay back slowly. A two-person team can run one slow channel and one fast one, rarely more.

How fast will it tell you something? Early on, the speed of the signal matters as much as its size. A channel that gives a clear answer in three weeks lets you try several ideas in a quarter; one that needs six months gets one try.

A useful habit comes from the book Traction and its Bullseye framework: list every channel that could plausibly work, run small and cheap tests on the most promising few, then focus on the one that shows results (Brian Balfour's summary). The point isn't the list. It's refusing to bet everything on the first channel that sounds good.

Step 3: Write the bet before you start

One sentence, written before launch: which stage this should move, by how much, by when, and how you'll see it. For example, "LinkedIn retargeting should lift demo requests about 20% above usual within six weeks, measured by demo requests with utm_source=linkedin".

That sentence protects you from the most common review trap, picking whichever number happened to go up afterwards. It also forces you to give the action its own signal: a campaign tag, a CRM source value, a dedicated form. The action log guide covers how to keep these lines together and review them weekly.

Step 4: Give it a fair run

Every channel has a natural delay between launch and effect, and judging too early produces "not working" verdicts on things that just haven't had time.

View full size

• Paid ads: Google advises looking at Smart Bidding performance over two conversion cycles or a two-week window, whichever is longer (Google Ads Help). For B2B, where a "conversion" may be a demo request, a month is a safer minimum.
• Outbound: one or two full reply cycles, including follow-ups.
• Events and webinars: until the follow-ups have been worked, not the day after.
• Content and SEO: months, not weeks. Judge early on leading signals like impressions and rankings, and on leads only later.
• Anything aimed at revenue: at least one typical sales cycle. Until then, judge it on the stage it moves first, such as meetings or opportunities.

If your numbers are small, a handful of leads a week, combine weeks and look at direction rather than single-week swings. "Directionally positive after six weeks" is a perfectly good interim call.

Step 5: Know when to look at competitors

Competitor research is useful at two moments and mostly a distraction in between.

Before you choose a channel. Where are they visible, what do they promise, and who are they talking to? Public ad libraries make this quick: LinkedIn's Ad Library, Meta's Ad Library and Google's Ads Transparency Center show what a company is running. A channel where nobody in your category is present can be an opportunity or a warning; a channel where everyone is present means you need a sharper message, not a bigger budget.

When a working channel stalls. If costs rise or replies drop without any change on your side, check whether a competitor changed its offer, entered your ad auctions or launched a cheaper plan. That's when their moves explain your numbers.

In between, checking competitors every week tends to produce copying, and copying gives you their positioning with a smaller budget.

Step 6: Keep, change or stop

When the fair run ends, the call is usually one of four.

Write the call down next to the bet. In six months, the record of what you stopped is as valuable as the record of what worked.

A one-page checklist

1. Which stage leaks most against its usual level?
2. Which channels can reach the buyers at that stage?
3. Of those, which can we afford and which will show a signal soonest?
4. What's the bet: stage, signal, target, deadline?
5. What's a fair run for this channel and our sales cycle?
6. Have we checked what competitors run there, and why?
7. On the deadline: keep, change, stop, or wait?

Common traps

• Choosing the channel first. "We should be on TikTok" is a channel looking for a problem.
• Spreading thin. Five channels at 20% effort each produce five unclear results.
• Judging on the first spike. Week two is rarely the answer; the fair run is.
• Scaling three things at once. If budget, audience and creative all change and results improve, you won't know which to keep.
• Following competitors into every channel. Their presence tells you where the attention is, not where your buyers decide.

Quick Q&A

How do I choose marketing channels for a B2B SaaS company?

Start from the funnel stage that leaks most, then pick channels where your buyers already look, that you can afford in money and time, and that will show a signal quickly. Test a few cheaply and focus on the one that works.

How long should I run a marketing campaign before judging it?

It depends on the channel: two to four weeks for paid ads, one or two reply cycles for outbound, until the follow-ups are worked for events, months for content and SEO, and at least one sales cycle for anything aimed at revenue.

How many marketing channels should a small team run?

Usually one or two at a time: one that shows results quickly and, if there's capacity, one slower channel that compounds, such as content.

When should I do competitor analysis?

Before choosing a channel, to see where competitors are visible and what they promise, and when a working channel stalls without a change on your side.

What should I do if a campaign is neither clearly working nor failing?

If it hasn't had a fair run, wait until a set date. If it has, combine weeks, run it longer or compare with a group that didn't see it before deciding.

Keep reading

• Did your campaign actually work?: baselines, fair test windows and holdouts.
• Track marketing actions, not just traffic: a simple action log, reviewed weekly.
• Why your converters show up as direct traffic: where the source gets lost.