Onboarding Drop-Off Versus the Enterprise Logo
Craftline is a B2B collaboration SaaS with 1,800 accounts. Only 38% of new accounts activate within 14 days; activation means creating a workspace, inviting two colleagues, and posting a document. Support tickets show 52% of new users abandon at workspace setup, where they must choose folder structures before seeing value. The product team of six has 12 onboarding improvements in the backlog. Meanwhile, a 4,000-employee prospect will sign a $600,000 annual contract if SSO is delivered in eight weeks; the sales VP calls it 'the only priority.' Only 2% of existing accounts have requested SSO, but they include two of the largest. The team is mid-way through an onboarding experiment and has no dedicated designer. The CEO wants a decision by Friday.
1. How should the product owner evaluate the SSO opportunity against the onboarding problem?
Read the worked answer
Compare the two on value, urgency, risk, and opportunity cost, not on who is loudest. The SSO deal is $600,000 annual recurring revenue and has an eight-week deadline, but it is one prospect; 38% activation affects every new account and may be depressing retention. Estimate cost of delay: for SSO, the contract value and risk of losing the logo; for onboarding, the lifetime value of the 62% who do not activate and the support cost. Check the sales claim: is the eight-week date a hard customer event or an internal target? Ask whether a limited SSO pilot-SAML for the prospect, no self-service admin-could satisfy the contract while the team continues onboarding. Size both: SSO may be 8-10 weeks for a six-person team, not a two-week add-on. Use a simple weighted model with strategic fit and risk, but make the trade-off explicit. Recommended decision: protect the onboarding experiment, offer the prospect a scoped SSO commitment with a named dependency and executive sponsor, and set a go/no-go review in two weeks.
2. What experiment or MVP would you run on onboarding before committing significant capacity? State the hypothesis, method, and success criteria.
Read the worked answer
Before committing eight weeks, test the onboarding hypothesis cheaply. The support data points to workspace setup and forced folder structure, so prototype a guided setup that creates a default workspace and defers folder choice until after the first document is posted. Run moderated usability tests with 6-8 newly signed-up users, including at least two administrators, and measure time to first document and completion without help. If traffic supports it, run an A/B test with activation as the primary metric and support contacts and workspace abandonment as guardrails. Define success in advance: for example, a 10-percentage-point lift in 14-day activation, or a 25% reduction in setup abandonment, without increasing support contacts. Time-box the experiment to two weeks and stop or iterate on the evidence. This is faster and cheaper than building a large onboarding redesign and gives the CEO a defensible basis for sequencing SSO.
3. Draft the recommendation you would give the CEO on Friday. How do you address the sales VP's deadline?
Read the worked answer
Give the CEO a one-page decision with three options. Option A: SSO first, onboarding paused; meets the prospect date if scope is limited, but leaves the activation problem untouched and consumes the team. Option B: onboarding experiment first, SSO after; improves the metric affecting all accounts but risks the $600,000 deal. Option C: split capacity-stabilize onboarding with a two-week experiment, then commit a scoped SSO with the prospect, provided the prospect accepts a pilot definition and the sales team owns the commercial risk. Recommend C, with conditions: the SSO scope is SAML for one identity provider and no self-service administration; the onboarding experiment runs first and reports at day 10; the sales VP confirms the customer will sign against the scoped definition. Show the cost of delay for each option and the assumption that the deal is lost if the date slips. This is a portfolio decision, not a backlog argument, and it needs the CEO to own the trade-off.