How Do You Build a Competitive Intelligence Practice That Sales Actually Uses?

How do you build a competitive intelligence practice sales actually uses?

Collect from a short list of public sources on a fixed schedule, distil it into one page per competitor written for a rep in a live call, and update it when something changes rather than quarterly. The failure is almost never collection. It is that nobody reads what was collected.

Every company we work with has some form of competitive intelligence. Almost none of it is used. It lives in a slide deck from eighteen months ago that a product marketer built once and nobody owns now.

Here is a framework for building the version that survives, and the public sources worth watching.

Why does most competitive intelligence fail?

It is written for the wrong reader. A thorough competitive analysis is written for leadership, to inform strategy. A battlecard is written for a rep who has ninety seconds and a prospect asking why they should not buy the other thing. These are different documents and most teams only write the first one.

It also decays silently. A competitor changes their pricing page and your document is now wrong, but nothing tells you. Reps discover the error in front of a prospect, lose confidence in the document, and stop opening it. That is usually a one-time event with permanent consequences.

And it is often about the wrong competitors. Teams track the two rivals they think about most, while deals are actually lost to an internal build, a spreadsheet, or doing nothing. Our notes on competitor comparison pages cover the same blind spot on the website side.

Which competitors should you actually track?

Start from lost deals, not from the market map. Pull the last fifty closed-lost opportunities and count what was chosen instead. The list you get is usually shorter and stranger than the one on your positioning slide.

Separate three categories, because they need different responses. Direct alternatives that do roughly what you do. Adjacent products that solve the problem differently. And the null option, which is the prospect doing nothing or keeping their current process.

The null option is nearly always your biggest competitor and almost never gets a battlecard. That is a real gap, because the objection handling for "we will just keep using spreadsheets" is completely different from the handling for a named rival.

What public sources are actually worth watching?

For public companies, regulatory filings are the richest source and the least used. The SEC describes EDGAR as providing "free public access to millions of informational documents filed by publicly traded companies and others," and its advanced search covers "the full text of electronic filings since 2001."

Full-text search over filings lets you find how a competitor describes its own risks, its segments and its strategy in a document it is legally accountable for. That is a very different register from their marketing site, and the gap between the two is often the most useful thing you will find all quarter.

For everyone else, the public surface is their own website, their changelog, their job postings, their documentation and their pricing page. Job postings in particular tell you what they are building six months before it ships, because they have to describe the work to hire for it.

How should you use search demand data?

Carefully, and with the caveats the source itself states. Google Trends explains that "each data point is divided by the total searches of the geography and time range it represents" and the result is "scaled on a range of 0 to 100 based on a topic's proportion to all searches on all topics."

So the numbers are relative, not absolute. A competitor at 100 and you at 20 does not mean they have five times your search volume in any countable sense. It means their share of searches in that window and geography was five times yours.

Google also states plainly that "only a sample of Google searches are used in Google Trends," that the data "incorporates statistical noise that includes small and random fluctuations," most noticeable on low-volume queries, and that "Google Trends is not a scientific poll." Use it for direction, never as a figure in a board deck. Our piece on branded search demand covers what to do with the signal.

What belongs on a battlecard?

Five things and nothing else. Where this competitor genuinely wins. Where you genuinely win. The two objections you hear most and the exact language that answers them. One proof point. The date it was last checked.

The honesty about where they win is what makes the rest credible. A battlecard that says your competitor has no strengths will be disbelieved by any rep who has lost to them, and disbelief is contagious across the whole document.

The date is the underrated field. A rep can calibrate how much to trust a claim if they know when it was verified. Without it, everything is equally suspect the moment one thing is found to be stale.

How do you keep it current without a full-time person?

Put the update on a schedule and make it small. One competitor per week, fifteen minutes, checking a fixed list: pricing page, changelog, job postings, review profile. Over a quarter that covers a dozen competitors properly.

Automate the watching, not the judging. Change detection on a handful of specific pages will tell you when something moved. What it means still needs a person, and that judgement is the part worth paying for.

Give it a single owner with a name. Competitive intelligence owned by a team is owned by nobody. It does not need to be a full-time role, but it needs to be someone's explicit responsibility.

What should you never do?

Misrepresent yourself to get information. Signing up for a competitor's trial under a false company name, or sending someone to a sales call pretending to be a buyer, is the line most companies eventually consider crossing when a deal is close.

Beyond the ethics, it is a bad trade. The information you get is marginal and the exposure if it surfaces is severe, particularly in a category where your buyers talk to each other. Everything genuinely useful is public.

Also avoid publishing anything about a competitor you cannot source. A comparison page with a claim you cannot back is a legal risk and a credibility risk at once, and competitors do read these pages.

How does this connect to how buyers actually decide?

It should shape when you intervene, not just what you say. The 6sense 2025 B2B Buyer Experience Report, based on nearly 4,000 responses plus a 766-response companion survey, found buyers purchase from their day one shortlist 95 percent of the time and evaluate an average of 5.1 vendors, with typical purchases involving ten or more people.

Read alongside the finding that first contact now happens at 61 percent of the journey, down from 69 percent in 2024, that points somewhere specific. Competitive positioning has to exist in public, before a rep is involved, because the shortlist forms without you.

Which means your battlecard and your comparison pages are the same content in two registers. If sales is saying something in calls that your site does not say anywhere, the buyer never hears it. Our notes on defining your ICP cover the input side of the same problem.

What should you build first?

One battlecard for the competitor you lose to most, and one for doing nothing. Two pages, written for a rep mid-call, with a date on each. Ship those before you build a system.

Then sit in three sales calls and watch whether they get used. That observation will tell you more about what to build next than any amount of internal planning, and it will usually reveal that the format is wrong rather than the content.

If you want to talk through how your competitive position should show up on your website rather than only in sales conversations, we are happy to walk through it. Come and find us at phoenix.studio.