[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"article-body-enterprise-software-has-about-twelve-problems":3},"\nWe make a provocative claim about enterprise software. Across the\nthirty-odd modules of a full suite,\n[roughly a dozen problem classes do almost all of the real work](/insights/twelve-problems-not-thirty-modules),\nand conventional suites rebuild each one per module, at the\ncustomer's expense. A claim with a count in it earns the obvious\nquestion. Name them.\n\nFair. A count you cannot enumerate is marketing, and we have\nwritten before about what separates\n[evidence from marketing](/insights/the-backtest-includes-the-misses).\nSo here is the list.\n\n## The dozen\n\n- **Entity resolution.** Is this supplier, this counterparty and\n  this hotline subject the same real-world thing, behind four\n  spellings and three registrations?\n- **Ownership tracing.** Who stands behind this company, layer by\n  layer, and when does indirect control or shared exposure matter?\n- **Forecasting.** What will this number be, and when will this\n  event happen, with stated odds and a naive baseline to beat?\n- **Anomaly detection.** Is this expense, override or sensor\n  reading unusual, against its own history and against its peers?\n- **Classification and mapping.** What kind of thing is this, and\n  what in the company does it touch?\n- **Conformance checking.** Does reality match the commitment? The\n  document against the changed rule, the actuals against the plan,\n  the transaction against the control.\n- **Exception triage.** Of these thousand signals, which few\n  deserve a person, at what severity, owned by whom?\n- **Scenario and counterfactual.** What happens if we act, or do\n  not, priced before the intervention is committed?\n- **Attribution.** Why did the number move? The answer arrives\n  ranked by measured driver, a calculation rather than a story.\n- **Allocation under scarcity.** Where does the scarce unit of\n  stock, cash or attention earn or protect the most?\n- **Planning and stability.** What sequence is feasible, and when\n  is a replan worth the churn it causes?\n- **Evidence and lineage.** Can every answer show where it came\n  from, and be regenerated identically on demand?\n\nNotice the nouns inside each question. Every one borrows from\nseveral functions at once, on purpose, because the nouns are\ninterchangeable. Swap supplier for customer, stock for cash,\nexpense for sensor reading, and the question does not change.\nThat is the definition of a problem class. A question shape, an\ninput shape and a way of scoring the answer that survives the move\nfrom one function to the next.\n\n## Maturing fields make exactly this move\n\nThe consolidation has a distinguished history. In 2006, Berkeley's\nparallel computing group published what became known as the\n[thirteen dwarfs](https://www2.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-183.html),\nthe observation that essentially all of scientific computing\nreduces to thirteen computational patterns. Build for the\npatterns rather than the applications, the argument ran, and a\ngeneration of hardware and libraries followed it. Operations\nresearch made the same move half a\ncentury earlier, when allocation, assignment and scheduling turned\nout to be the same problems wearing different industrial costumes,\nwhich is why one method conquered refineries, airlines and\nlogistics at once.\n\nEnterprise software has not made the move, and the market's own\nstructure shows the cost. Entity resolution is sold today as\nmaster data management, KYC matching, sanctions screening and CRM\ndedupe, several of them with their own analyst categories.\nAnomaly detection is sold as fraud detection, transaction\nmonitoring, quality control and expense audit. The industry\nmaintains separate product categories for the same underlying\ncapability with different nouns attached, and every customer funds\nthe duplication.\n\n## Run the test on the list\n\nA published list is an invitation to check it, which is the point.\nTake any capability of any module, ours or a competitor's, and ask\nwhich of the dozen it reduces to. Supplier risk warnings reduce to\nforecasting plus exception triage. Regulatory change management\nreduces to classification and mapping plus conformance checking.\nSanctions exposure reduces to entity resolution plus ownership\ntracing. If you find a real capability that does not reduce, the\nlist grows, and we would rather grow it in public than defend a\nnumber.\n\nThe count is honest at the family level. Inside each class sit\nspecialisms, the intermittent series that needs different\ntreatment from the steady one, the cold start that borrows from\nan analogue. Those are depths within a class, not new classes,\nin the same way cardiology and paediatrics are depths within\nmedicine.\n\nThe important argument is not whether the number is twelve. It is\nthat the number is small.\n\nEnterprise software inherited the organisation chart as its\narchitecture. Finance bought finance software, compliance bought\ncompliance software, supply chain bought supply chain software,\nand the result is a market that repeatedly solves the same\nunderlying problems under different names. A maturing industry\neventually notices that it is rebuilding the same machinery.\nComputing noticed in 2006. Operations research noticed in the\n1950s. Enterprise software is beginning to.\n\nSo the question is no longer whether these problem classes exist.\nThe question is whether organisations want to keep paying to\nsolve them thirty times. Prophesee was built on the belief that\nthey should be solved once, improved continuously, and shared\neverywhere they apply. Ask a suite vendor to name their dozen,\nand watch whether the answer is a list or a subject change.\n\nBuild around functions and every improvement stays local. Build\naround problems and every improvement compounds.\n[See the classes running on your own functions](/contact).\n",1786984935416]