why 70% of digital transformations fail
think metamorphosis, not transformation!
New thought, not just new technologyDigital is a game-changer but only if you’re prepared to change your thinking; understand why the status quo is flawed & how to get back on track!
Certain technologies compromise long-term architectural foundations.
Data growth is exponential, so associated issues & costs also grow exponentially!
Information architectures aren’t purpose-built. Tactical short-term fixes create long-term problems & increase technical debt.
Governance is application-limited rather than systemic.
The people, processes & technologies within your ecosystem are incoherent. They all ‘speak’ different languages.
Given this context, the wonder is that any digital transformation projects succeed!
Data growth is exponential, so associated issues & costs also grow exponentially!
Information architectures aren’t purpose-built. Tactical short-term fixes create long-term problems & increase technical debt.
Governance is application-limited rather than systemic.
The people, processes & technologies within your ecosystem are incoherent. They all ‘speak’ different languages.
Given this context, the wonder is that any digital transformation projects succeed!
Digital should be data-centric!
Why data-centric?Being data-centric means putting data (not applications) at the heart of your digital domain.
Every application has its own database, which exacerbates volume & complexity, increasing cost while diminishing value.
Every database is different & comes with its own data model. For instance, a vanilla SAP implementation has over 90,000 tables, each of which has several columns so the numbers get very large very quickly.
Every time one of your processes needs data, that means integration. Every integration is a cost & a time sink.
Research shows that organisations typically use only 20% of the data they hold, therefore destroying the value proposition as they pay for data they don’t use, don’t need and maybe, shouldn’t hold.
Every database is different & comes with its own data model. For instance, a vanilla SAP implementation has over 90,000 tables, each of which has several columns so the numbers get very large very quickly.
Every time one of your processes needs data, that means integration. Every integration is a cost & a time sink.
Research shows that organisations typically use only 20% of the data they hold, therefore destroying the value proposition as they pay for data they don’t use, don’t need and maybe, shouldn’t hold.
Realise the value of data.
to fix digital, fix the datadata should be:
is data giving you a corporate coronary?
Digital is only as good as its dataIn an era of extreme data, the aim is to reduce volume & complexity otherwise the result is too much data – too little value.
Data must be valid for decisions based on it to be sound.
Data should be exposable, theoretically open to every process. Bezos thought this principle so important he made failure to do so grounds for dismissal.
Data that is not described (i.e. without metadata) is virtually useless. The system will be unable to find it or know what it represents. Organisations typically have 80% dark data: cost-generating, meaningless digital real estate instead of valuable knowledge assets.
Data that is not properly formatted (e.g. Rd. /road /Road) slows up the system; thwarts search functions; & demands avoidable time & effort be spent on parsing corrections.
Data should ideally be discrete otherwise it’s a cost multiplier.
Data should be exposable, theoretically open to every process. Bezos thought this principle so important he made failure to do so grounds for dismissal.
Data that is not described (i.e. without metadata) is virtually useless. The system will be unable to find it or know what it represents. Organisations typically have 80% dark data: cost-generating, meaningless digital real estate instead of valuable knowledge assets.
Data that is not properly formatted (e.g. Rd. /road /Road) slows up the system; thwarts search functions; & demands avoidable time & effort be spent on parsing corrections.
Data should ideally be discrete otherwise it’s a cost multiplier.
What is the cost of doing nothing?
From sclerotic to optimal processesProcess should be a multiplier
Like search engines, processes can’t function without metadata! Application opacity makes it very difficult to wrest meaningful metadata from application control, causing:
It shouldn’t be this hard!
Process inefficiencies are eminently solvable if the metadata is good.
Process inefficiencies are eminently solvable if the metadata is good.
AI can only infer metadata from pre-existing patterns
to fix process, fix the metadata
don’t use AI to do the housekeeping!
The ‘google’ in your backyard!Systems are supposed to be a coherent whole. Systems speak the language of applications, machine metadata, but without semantics, human metadata, man & machine can’t talk to each other. Your systems aren’t actually….. systems.
Unless there is consistency from the boardroom to the edge, your ecosystem will be incoherent. You will continue to reap cost instead of value.
To fulfil that definition, your ecosystem needs an organising mind, an ontology, so that there is domain-wide consensus over entities, concepts & meaning, informed by context & relationships.
Without ‘knowing’ your digital estate, your ability to leverage AI, to understand the present & consequently infer the future, is compromised.
Information has to be coherent & consistent across the domain: from the edge to the centre; from end user to CEO.
Scale up your perspective! That ‘system’ in your backyard, that’s your own private ‘Google’!
How will you ‘google’ your own domain?
to fix systems, be coherent
no data, no intel
no knowledge, no AI!
no knowledge, no AI!
if you know you can grow
how will you reach the stars?
Evolutionary advantageThe use of computers, just like the ‘horseless carriage’, enables man to go further, faster & with less manual effort; to scale, perform & control.
Knowledge enables human ideation (the growth mindset) to leverage compute power. Tools such as AI, offer supra-human capabilities – evolutionary advantage, but only if man & machine communicate. Optimisation needs coherence: to gain new knowledge, future insights or enhanced capability through applying probabilistic, automated reasoning to existing knowledge assets.
From raw data, analysis informs the business trajectory by applying human ‘intelligence’. Intelligence is not something that can be ‘artificially’ imposed. True ‘intelligence’ is the value of what the ‘horseless carriage’ enables humans to do, not the ‘carriage’ itself! Intelligence evolves organically e.g. by capturing & leveraging existing IP, skills & expertise – & then reapplying them using AI.
By understanding where we’ve come from & knowing where we are; we can better use the predictive capability of AI to infer potential scenarios from which we can select the optimal direction of travel.
From raw data, analysis informs the business trajectory by applying human ‘intelligence’. Intelligence is not something that can be ‘artificially’ imposed. True ‘intelligence’ is the value of what the ‘horseless carriage’ enables humans to do, not the ‘carriage’ itself! Intelligence evolves organically e.g. by capturing & leveraging existing IP, skills & expertise – & then reapplying them using AI.
By understanding where we’ve come from & knowing where we are; we can better use the predictive capability of AI to infer potential scenarios from which we can select the optimal direction of travel.