First the hint (in What next after part I?):

I am already able to demonstrate the automatic identification of survey scores that don’t fit the model – areas in which the organisation in question has unusually strong areas to celebrate or weak areas for further investigation – all driven by the data, not the biases of the person facilitating the debrief

Now the announcement: the Agendashift unbenchmarking service is now available.

In a nutshell: we run your survey results against our dataset and identify areas of opportunity for you automatically.

When and why to unbenchmark

• You’re embarking on a journey of Lean-Agile transformation and you want a better understanding of the landscape you’ll be navigating
• Your improvement efforts are stalling and it’s time to agree priorities afresh
• As input to your corporate culture exercise, team retrospective, departmental meeting, or offsite
• Anytime you need to reorient, re-baseline, take stock, and re-energise
• When it’s time to take a regular checkpoint
• As a low-commitment step before taking an Agendashift workshop or engaging an Agendashift partner

How it works

• We set up for you an online survey with the full 43-prompt Agendashift values-based delivery assessment
• You distribute it internally for your staff to complete by the agreed deadline
• We report back to you:
  • The overall distribution of your scores
  • Stronger and weaker categories and prompts, as measured in absolute terms (without reference to other data)
  • Categories and prompts whose aggregate scores are above or below the expected profile given the correlations we have observed
  • Areas of agreement and disagreement, reinforcing the more obvious findings and creating opportunities for story-gathering where the picture is more mixed
  • The complete set of aggregate results for all categories and prompts (twice over in fact – one set for prioritisation purposes, another for planning)

If you’re a veteran of one of our workshops or you have read the Agendashift book (chapters 2 and 3 in particular), you’ll recognise this reporting structure. It reflects how we typically organise a survey debrief and the mapping exercises that follow.

Unbenchmarking

The new and exciting bit is the identification of data points above or below the expected profile, ie scores that differ from what might be expected given your survey’s overall shape and the correlations we observe elsewhere. We’ve used machine learning techniques on our growing dataset to create a model of survey results to which yours can be fitted; the deviations above and below the best-fit model identify what’s particular to you. Instead of a crude index or vanity metric that’s meaningless in the context of an open-ended journey of transformation, we give you detailed insights you can actually work with.

Additionally, we can control the granularity of the model so that you see deviations not just from the total population but from organisations that show traits similar to yours. This helps you put your findings into proper context, and makes the insights even more actionable.

Should you decide to repeat the exercise at some later date, any significant changes to your profile will cause new areas of interest to be highlighted. Just as you’d expect!

In this real example, the three prompts whose scores lie furthest below the expected profile clearly have something in common:

As illustrated here, the biggest opportunities can hide in unexpected places, away from where absolute scores are obviously weak. Unbenchmarking highlights these potential action items for you automatically.

What happens next is of course up to you. However, the process of turning prioritised prompts into an agreed plan of actionable outcomes appropriate to your unique context is something we are well practised at facilitating; no matter wherever you are located, we should be able to find you some valuable support.

Pricing

The Agendashift unbenchmarking service is priced at £125 (about €145 or $160) per survey plus £1 per person over 10 people surveyed, with survey sizes specified in advance.

Several discounts are available:

• We make generous allowances for charitable, non-profit, educational, and government organisations
• If you’re based outside the major economies we can make PPP adjustments
• For large exercises involving multiple teams or departments, you may find it helpful to receive both separate and combined results; pricing will depend on complexity but multiple surveys will in general attract discounts.
• Similarly, repeat exercises (eg quarterly) will be discounted

The Agendashift values-based delivery assessment is already included as prework for our transformation mapping and practitioner’s workshops. Where workshops follow unbenchmarking we’ll adjust workshop pricing appropriately. You won’t pay twice!

Finally, you should consider engaging one of our partners, professionals well qualified to support you in your journey of transformation.

Get in touch now to arrange your unbenchmarking exercise:

Arrange an unbenchmarking exercise

Related

• Which assessment template?
• Which workshop?
• The Agendashift partner directory
• The Agendashift partner programme
• The book: Agendashift: clean conversations, coherent collaborations, continuous transformation, chapters 2 and 3 especially

Meet us online

• Read our blog: blog.agendashift.com
• Join our Slack community: agendashift
• Join our LinkedIn group: Agendashift
• Follow us on Twitter: @agendashift

What’s next after part I of the new Agendashift book (published last month), that is.

The obvious answer would be part II, which I intend to consist mainly of case studies and smaller real-world examples of Lean-Agile change in action. Before that though, I’ve been experimenting with some machine learning tools, in particular in the area of dimension reduction. Broadly, this allows the 43-dimensional data of the Agendashift survey results to be boiled down to just a few key features (factors, traits, tendencies, etc).

This has considerable potential:

1. I am already able to demonstrate the automatic identification of survey scores that don’t fit the model – areas in which the organisation in question has unusually strong areas to celebrate or weak areas for further investigation – all driven by the data, not the biases of the person facilitating the debrief
2. Cluster analysis could allow representatives of organisations exhibiting similar traits to be introduced and their experiences compared
3. A combination of 1 & 2 could lead to some kind of recommendation system – machine-assistance for the coach, perhaps

Expect an announcement very soon on point 1, a new commercial offering and new functionality available soon to paid users (ie Agendashift partners).

Point 2 is in fact my plan for part II. Technicalities aside: bring people together, get them talking, write it up 🙂

Point 3 is for the longer term. A “living” version of the book, if you like.

Technical footnote: The machine learning (ML) tools available in Python are truly awesome. Book recommendation: Python Machine Learning by Sebastian Raschka. I’ve made good use of his online tutorials too.

Questions? Go to #next-steps in our Slack.