adoption

In depth

Four states

Four states are commonly confused, when telling them apart is enough to clear up most discussions. A tool that is rolled out is available; a tool that has been tried has been opened once; a tool that is used comes back on a real task; a tool that is adopted would be missed if it were taken away. The measure follows from that scale: neither licences nor sign-ins, but recurrence on real work and the share of the activity concerned. The most instructive signal is the return after the first failure, because an abandoned tool is almost always a tool that disappointed once without offering a way back. Adoption only arises, moreover, where a person has to change a habit, which separates it clearly from automation, where the task disappears without anyone having to take it up.

What makes a use stick

Three conditions recur in the uses that settle in for good. The tool arrives where the work is done rather than beside it, which avoids adding one more switch to an already fragmented day. The gain is perceptible from the first attempts, in the person’s frame of reference and not in the project’s, failing which the effort of learning is granted out of obedience and stops with the attention paid to the subject. The cost of checking stays below the gain, which is the condition most often ignored and the most decisive. Finally, adoption is never uniform across a team: a minority of intensive users is enough to produce a flattering average, which makes reading by segment more honest than reading by average.

What blocks

What blocks adoption rarely lies in the tool alone. The absence of an approved equivalent leads everyone to equip themselves on their own, which produces uses outside the rules rather than an absence of use. An unstated rule has a similar effect, because in the absence of a clear answer the cautious abstain and the others go ahead without saying so, which distorts any measure. Conversely, adoption made compulsory produces compliance with an instruction: measured usage rises, the effect on the work stays nil, and the indicator becomes misleading at the precise moment one believes it has been improved. Adoption must therefore be kept in its place, that of an intermediate indicator: what is sought is not that the tool be used, but that the work be changed by it.