What It Looks Like
This chapter departs from the composite examples used elsewhere. It is a single real engagement, anonymised and compressed; the note in the appendix explains how, and what it is and is not offered as evidence of.
The programme was recorded as an IT failure because that was the only category large enough, and safe enough, to receive it.
A multi-country asset-services group had spent years preparing to replace and standardise the systems through which its national operating entities ran their core business. The plan was familiar. Select a global enterprise platform, appoint a large systems integrator, configure the product, roll it out country by country, and absorb local variation through design decisions, change requests, and governance. On paper the logic was defensible: one brand, recognisable services, countries that looked similar from the altitude at which business cases are written, and a technology estate that was fragmented, expensive, and heavily customised. A standard platform promised standardisation, and standardisation promised speed.
The promise failed for a reason the programme structure was not built to see. The countries were not variants of one explicit business but locally coherent businesses using overlapping language. The same words did not mean the same things, the same customer promise did not imply the same operational process, and a service that looked standard in the commercial vocabulary decomposed differently the moment it touched suppliers, insurance, contract events, billing, customer communication, and local regulation. The organisation had set out to standardise systems before it had standardised meaning.
The software did not fail to implement the business. The business had not yet become implementable.
The group looked standard from above. It operated under one brand across several national entities, sold recognisable services to recognisable customers, ran a similar financial model, and shared a broad customer lifecycle from contract origination through service to termination. Viewed through the categories in which executives and vendors describe a business, brand, sector, revenue, customer type, rollout schedule, it was one business with an expensive technology problem.
The categories were the problem. The enacted business lived below them, in the promises each country actually made and the processes that fulfilled them, and at that level the entities had diverged for years. They used the same words for different commitments. They had different service offers, supplier arrangements, regulatory habits, operational exceptions, and local system behaviours. Terms that appeared common at head-office level decomposed differently in practice, and no one had been required to notice, because nothing forced the words to mean the same thing.
A restaurant group makes the same mistake in miniature. Head office believes it owns a standard burger chain because every outlet shares the brand, so it procures a standard burger operating system: one point-of-sale model, one kitchen workflow, one inventory structure, one training programme, one reporting taxonomy. Only during rollout does it discover that some outlets sell beef burgers, some sell chicken wraps, some sell tacos, some serve a halal menu, and several use the word “meal deal” for different bundles with different substitutions, price rules, kitchen steps, supplier constraints, and allergen controls. The failure is not that head office wanted standardisation. The failure is that it tried to standardise the operating system before it had discovered the menu.
...
Continue reading in the interactive reader
Read this chapterSee also: Full contents · Preview chapters · Illusions in the Boardroom