Why "We improved throughput by 30%" loses the room

Platform leaders who lead with Jira velocity, DORA, or SPACE scores invite the wrong conversation, according to Christoph Steinlehner.

"Both measurements are important, but on their own, they are poor tools for communicating outside the team," he writes. "They work within the team to optimize, but as soon as they are used externally, they invite the wrong conversation."

Two failure modes appear repeatedly. Teams optimize purely for developer experience and wonder why leadership won't fund them. Others look only at the business side and never solve the right consumer problem.

"Holding both at once is the core tension in platform product management," Steinlehner says.

What does the four-driver model actually look like?

Steinlehner borrows a frame from Joshua Arnold and John Cutler that separates business KPIs by direction:

• Increase revenue
• Protect revenue
• Reduce costs
• Protect costs

Splitting protection into revenue and costs aligns typical platform work — governance, compliance, reliability — with board-level KPIs. One driver often applies; sometimes two or three stack in the same chain.

How do you trace a consumer action to a business KPI?

The logic chain: consumer action → team-level outcome → business KPI → business impact. The middle steps vary by org size, but the link stays intact.

The wrong way: define both ends, then claim a connection. The right way: build the chain together, then test it.

Steinlehner walks through a payment platform example. Multiple teams had integrated payment handling, duplicating maintenance. Centralizing removed duplicate work and enabled faster rollout of new payment providers. That lifted checkout conversion and revenue. The platform team's job ended at adoption; conversion lived with the consuming teams.

A second example: leadership mandated AI usage across roles. Support wanted case summaries; account managers wanted customer-history insights. Raw customer data couldn't enter an AI system without access controls and classification. A secure, internal AI platform turned the mandate into reality and protected revenue through compliant service and lower churn.

How concrete should a consumer goal be?

Very. "Account managers reduce the time to get a full account overview from 10 minutes to below one minute" beats "account managers get faster."

He frames the math bluntly: an organization spending €50k per sprint on a platform team that returns €200k should not be having a cost-cutting conversation. The right conversation is what to invest in next.

What changes when the team speaks this language?

Budget talks shift from negotiation to prioritization. Steinlehner: "Instead of constantly negotiating spend, this approach shifts to what's worth investing in."

Prioritization becomes defensible. Most scoring systems include value or ROI ratings that devolve into politics — every initiative arrives at 9-out-of-10. Grounded consumer-to-business logic replaces guesswork with testable behavior change.

Leadership gains leverage. A leader saying, "Our platform team supported €X in revenue growth by providing Y for Z" sits in a different position than one citing deployment frequency gains of 300%.

What about adoption metrics?

Adoption is a useful starting point. Optimizing for it alone is "like optimizing for revenue" — abstract, and open to any random idea. Mandated adoption is meaningless because consuming teams have no choice.

Where does this head next?

Platform product management is moving from delivery talk — "we improved throughput by 30%" — to enablement talk — "we enabled four teams to deliver compliant AI features, lifting enterprise contracts by 5%." As internal platforms compete with vendor SaaS for budget, the teams that can cite a 4x return on platform spend will out-argue those still presenting Jira dashboards to the CFO.