Why CI tools become shelfware: six adoption signals to check by week four

Six numbers, written before the data

In the first month a PMM or CI manager is asked whether sales is using the tool. The harder question is what "using" means and what it is divided by. The pages that come up for that search are written by vendors, and none of them answers it.

The nearest thing to a vendor definition is Klue's own Consumer Usage dashboard, described as "the % of your content consumers who access your program over time" (Klue, read 7 October 2026). That has a numerator and a denominator. The quoted description mentions no window, and its denominator is everyone with access.

So define the measures yourself before you look at any number. For each of the six signals below, write the formula, the window and the value that would satisfy you, with the date you wrote it. The pass line is yours. This article proposes none, because no published benchmark for CI adoption with a stated denominator exists, and the all-SaaS figures in the next section are about licences in general. A line written after the data arrives is a rationalisation, whatever it says.

What the published numbers say

The shelfware figures you have read come from SaaS management vendors and cover every application a company buys, with no competitive intelligence subset. Two are worth knowing, for the form of the metric rather than the value.

Zylo's 2023 SaaS Management Index reported that "on average, 44% of businesses' SaaS licenses are wasted or underutilized", and that 70% of SaaS contracts were renewed in 2023 despite being underutilized (Zylo, 4 April 2023; sample and method are not stated on the page). Productiv's analysis of nearly 100 million licences over three years found that "only 47% of SaaS licenses are used over a 90-day period on average" (Help Net Security, 4 July 2023).

Productiv's figure uses a 90-day window; Zylo's page does not state its method. The useful part is the form: a window, a seat count and a count of seats with at least one event. The 44% and 47% are not targets for a CI tool, and the two vendors measure different things in different years, so do not add them together.

The one CI-specific ratio we found comes from Crayon, which divides total battlecard views by the number of competitive deals in a period. For one customer's top five competitors over 90 days, the figures included 1.92 views per deal (125 views over 65 deals) to 3.63 (69 over 19) (Crayon, 28 September 2023). Crayon reads a high value as a sign of low seller confidence rather than success. A cleanly defined ratio can still leave you unsure whether the program is working, which is why the six signals below are read together.

The six signals

Each signal below is this site's proposal. Each has a formula, and none has a threshold. Where a published source defines the form of the measure, it is named.

• Reach: distinct sellers who opened any card in the last 28 days, divided by provisioned seller seats. Your pass line, dated, before the first export.
• Return: of the sellers who opened a card in weeks 1 and 2, the share who opened one again in weeks 3 and 4.
• In-deal use: card views by sellers on opportunities with a competitor recorded, divided by competitive opportunities in the period.
• Where it is used: share of events that came from the seller's own surface (CRM, Slack, Teams, call tool) rather than the vendor's web app.
• Feedback loop: sellers who submitted at least one correction or question, divided by sellers reached; and the median days from a submission to a visible answer.
• Card quality: of the cards sellers actually opened, the share with a talk track and a dated proof point.

Reach is the licence-utilisation measure applied to sellers only. Admins, the CI team and the vendor's customer success manager open cards for reasons unrelated to selling, so they come out of both numerator and denominator. Twenty-eight days is a choice; what matters is that the window is fixed and written down.

Return separates curiosity from habit. A launch announcement and a kickoff demo produce a first open from most of the team. The sellers who come back in the second fortnight without being asked are the ones using it in deals.

In-deal use is Crayon's views-per-competitive-deal ratio with the denominator restricted to deals that have a competitor recorded. It is the signal that depends on the CRM, and the one most often blank. If your opportunities carry no competitor field, you cannot compute it, and that is the first fix in the next section.

Where it is used tells you whether sellers reach for the content at the moment of the deal or go looking for it in a separate application. Crayon's 2025 survey write-up describes the best programs as "shifting from creating documents to creating access" (Crayon, 5 August 2025). Treat that as a direction, since no sample size is given, and measure your own split.

The feedback loop has two numbers because a channel with submissions and no answers dies in a month. The population question, what share of sellers have ever used the channel, belongs here. What happens to one seller's correction once it arrives is covered in turning seller feedback into a reviewed battlecard revision.

Card quality has a published audit figure, and it is a vendor's. After "over 150+ battlecard audits", Klue reported that "only 43% include talk tracks to use" (Klue, 27 April 2023). Klue's audit also reports 35% with customer-facing proof points; we quote the talk-track figure as the one our signal uses. Audit the cards your sellers opened rather than the whole library; a card nobody opens can wait. A three-line battlecard format describes what a card with a talk track and a dated proof point looks like.

A signal you cannot compute is reported as unmeasured. Unmeasured is a finding about the tool or the CRM; zero is a finding about the sellers. Confusing the two is how a renewal conversation goes wrong in both directions.

Make it measurable in week zero

Most of the week-four scorecard is decided before anyone logs in. Four things have to exist on go-live day, and the vendor evaluation is the place to confirm them: test each tool with real sellers on your own competitors and deals, against pass/fail criteria agreed in writing first.

1. Write the seller list, with a date

   Names and roles of the people who sell, as of go-live. Everyone else who has a seat is excluded from reach and return. Keep the list; people join and leave during the four weeks.

2. Record the provisioned seat count

   The number of seller seats the contract provides, which is the denominator for reach. If seats are invited rather than provisioned, record invitations sent and accepted separately.

3. Add a competitor field to opportunities

   In-deal use needs the CRM to say which deals are competitive. Crayon's 2025 State of Competitive Intelligence page reports that "44% of companies lack competitor visibility within their CRM" (Crayon, read 7 October 2026; sample not stated). A picklist on the opportunity, filled by the seller or suggested by a tool and approved, is enough.

4. Get a sample per-seller export before you sign

   Ask each vendor for an export that names the seller, the card, the event and the timestamp, and ask what the delay is. A sample file is the proof, not a promise.

What is public about the three best-known tools, as of 7 October 2026, is thin. A third-party connector document describes a Klue API that returns "user event activity logs (card views, clicks, actions)", "updated daily at 00:00 UTC with a 3-day delay", with keys created by Klue admins (Nexla); Klue's own documentation was not reachable, so whether this is available on every plan is unknown. Crayon's seller engagement view combines battlecard views with win rates imported from Salesforce (Crayon, 14 December 2023); Crayon's seller engagement page does not state an export format. Kompyte advertises "behavior analytics and battlecard analytics" (Semrush knowledge base); export is not described. Each gap is a question for the vendor, and one that cannot answer it has given you signal zero.

The same applies to a program with no vendor. A Claude Project full of competitor files that nobody opens is shelfware with a lower invoice. A shared sheet where sellers log which card they used on which call, or a link tracker you control, gives you reach, in-deal use and the feedback loop by hand.

Read the result

A failing signal points at a cause, and Klue's diagnostic for low content adoption gives a usable frame. It asks four questions: do people know the content exists, do they know where to find it, is it relevant, and is it applicable (Klue, 3 April 2023). Mapped onto the six signals:

• Reach fails and return is unmeasured: awareness or accessibility. Sellers have not been shown the content, or the login is one more place to go.
• Reach passes and return fails: relevance. People looked once and found nothing about the deals they were in.
• Reach passes and in-deal use is low: accessibility at the moment it matters. The card exists but is not on the opportunity or in the channel where the call is prepared.
• Feedback loop at zero submissions: nobody expects an answer, or nobody has been told the channel exists. Check the median-days figure before concluding the sellers are the problem.
• Card quality fails: applicability. A card with no talk track gives a seller nothing to say.

The arithmetic is where Claude earns a narrow place. Attach the per-seller export and the seller list to a chat, or to the Project you already keep for competitive work, and ask for the six ratios with the counting steps shown and a list of rows it could not classify. CSV is a supported upload type, and Claude extracts text only from a CSV (Anthropic support, read 7 October 2026), so filter a large export to the window first. Then pick two sellers and re-count their events by hand. If the two match, trust the rest for this month; if they do not, you misread a column, and you found it before the renewal meeting. No prompt wording adds anything here; the instruction is the six formulas above.

What usage cannot tell you

Reach is not impact. A seller can open the card before every competitive call and say nothing different on it. Whether the card changed what was said is a question for call recordings and the feedback channel; whether it changed outcomes is a question for win rate and deal size, a separate measurement with its own traps, not covered here.

Four weeks is also shorter than many sales cycles. A seller with no competitive opportunity in the window has not failed in-deal use; the signal is unmeasured for that seller, and the scorecard should say so by name. Four weeks is this article's framing, long enough to see return and short enough to act before the first quarterly review. Nothing in the sources supports it as a standard.

Two more things escape every tool. A card used from a screenshot or a forwarded PDF is invisible to the vendor's analytics, so a low surface split may mean sellers found a shortcut rather than that they ignored the content. And none of the six signals says whether the card was right. A frequently opened card with a stale price on it is a liability that your usage data will describe as a success.

Keep the pass lines and the week-four export. They are the baseline the renewal conversation will be argued from.

Where Steve stands

What Steve's site states, as of 7 October 2026, covers some of the six signals and not others. On delivery, the competitive intelligence page says Steve sends competitor news summaries to the channels you choose every morning, and that reps can ask competitive questions in chat and get the relevant battlecards back; Slack updates can be daily or weekly, and a Zoom Team Chat bot answers the /battle-cards command. On the deal, Steve puts a battle-card link on the HubSpot deal or Salesforce opportunity and the rep's Slack or Zoom /prep brief names the expected competitor before the call, and it shows how often each deal's card was opened. It also shows win rate and deal size on deals where Steve was used versus not, per competitor. Updates to a card wait for your review before they apply. Intelligent Deal Detection is part of Team plans, not Starter.

That covers the surface signal and the deal-level part of in-deal use. Per-seller reach, return and a usage export are not stated on the site. Ask any vendor for the sample export in week zero, and write your pass lines before you open it. Steve competitive intelligence is the page to start from.