Thread: I've been playing around with some equations to assess whether it's worth automating a recurring process or not.
Let's start with X.
X = Baseline Cost in Hours to Run Process
That's how long it takes to do something once. Getting bookkeeping current, doing a marketing campaign, having a coordination meeting, whatever.
X = Baseline Cost in Hours to Run Process
That's how long it takes to do something once. Getting bookkeeping current, doing a marketing campaign, having a coordination meeting, whatever.
X might be possible to improve, or it might not. You'd have to do analysis to see if you could streamline the process or not.
We could represent that as such —
Analysis (Hours) * Chance-of-Successful-Realization (%) = Cost Alpha, α
That's the first cost.
We could represent that as such —
Analysis (Hours) * Chance-of-Successful-Realization (%) = Cost Alpha, α
That's the first cost.
If possible to automate, it's probably some mix of Definition, Standardization, and Building out tech and/or collateral. I break it out like that since it's easier to estimate in pieces than a whole block.
Definition + Standardization + Tech/Collateral Creation = Cost Beta, β
Definition + Standardization + Tech/Collateral Creation = Cost Beta, β
Of course, if it's not full automation immediately, you'll need to have some training and installation. That takes time too.
Training + Installation = Cost Gamma, γ
Training + Installation = Cost Gamma, γ
And... we're almost done.
[Cost of (α + β + γ) - (X * frequency-of-X-over-timeframe)] = Delta, δ
That’s hours saved per run of the process over a given timeframe. If the process happens once per week and you're checking the one-year payback, then frequency-of-X is 52.
[Cost of (α + β + γ) - (X * frequency-of-X-over-timeframe)] = Delta, δ
That’s hours saved per run of the process over a given timeframe. If the process happens once per week and you're checking the one-year payback, then frequency-of-X is 52.
And finally, let's not neglect that processes typically require some maintenance to keep running smoothly.
Ongoing Maintenance Costs of Process (Hours) = Cost Episilon, ε
(Over a given timeframe, obviously.)
Ongoing Maintenance Costs of Process (Hours) = Cost Episilon, ε
(Over a given timeframe, obviously.)
Annnnnnd, you ready for some magic?
If,
α + β + γ < δ - ε
over target time period, then it’s time-profitable to do.
Huzzah, eh? Ain't that beautiful?
If,
α + β + γ < δ - ε
over target time period, then it’s time-profitable to do.
Huzzah, eh? Ain't that beautiful?
Of course, feel free to replace "α" with "a", "β" with "b", etc.
If,
a + b + c < c - d,
then time profitable.
I'm just being fancy + Greek letters synchronized surprisingly well with the subject matter. The delta being the time-savings variable, epsilon being maintenance, etc.
If,
a + b + c < c - d,
then time profitable.
I'm just being fancy + Greek letters synchronized surprisingly well with the subject matter. The delta being the time-savings variable, epsilon being maintenance, etc.
I wrote a series of essays that covered ops and automation a bit that was popular if you're interested in a longer-form treatment, this was the first —
medium.com/the-strategic-…
medium.com/the-strategic-…
But yeah, I'm super hyped for this. Going to make a lightweight spreadsheet to analyze everything at Ultraworking, we're about to go on a big process improvement / efficiency run.
