Differentiation Without a Framework: Engineering Investment Decisions in Berlin Early Stage Startups

Multi-case qualitative analysis of 33 retrospective essays written by practitioners in six early-stage startup teams in Berlin who built real software products over three months. A deductive codebook derived from the Service Maturity Framework's constructs (Optimize/Constrain/Defer classification) was applied to extract five dimensions: differentiation behavior, driver language, translation gap evidence, uniform-threshold behavior, and MVP debt signals. Three findings: (1) differentiation is universal (100% of teams) but systematic differentiation is nearly absent (93.9% exhibit the translation gap); (2) teams naturally reason in terms the framework predicts — value/criticality (100%) and evolution stage (87.9%) — but 87.9% also cite unpredicted drivers, led by demo/visibility pressure (42.4%); (3) intra-team convergence on investment postures reaches 73% without formal training or a shared protocol, with agreement strongest at classification extremes and divergence concentrated in ambiguous middle cases.