📝

Harnessing Customer Feedback & Creating a "Feedback Machine"

  • Lots of different feedback (NPS, Csat, Trustpilot, etc.)
  • Goal: one place where all the feedback goes to
  • One owner for each source of feedback
  • Make sure that afterwards there’s a clear feedback route; but problem is how to do that at scale?
  • How do we no overengineer something that should be really simple?
  • How do we build something that works for all teams?
  • “Where we are with feedback / where our vision is?” > how do we build something that’s scalable?
  • Onfido: NPS: top level company metric. Vision: get feedback from all the sources and aggregate; split the data and analyse by all elements that we want
  • Onfido, Wagestream > B2B2C
  • Feedback gathering: ad hoc or purposeful basis?
  • Gathering feedback process nailed; then analysing and actioning on that feedback (how to automate it?)
  • Thread: Report NPS weekly to the whole company! Back it with the customer support tags!
  • is it possible to understand what the customer journey will look like off the back of the customer feedback?
  • NPS: great metric but potentially a lagging metric
  • How do you move from a resource intensive gathering of customer feedback to something a bit more scalable
  • E.g. the % of bookings that result in a ticket needs to be reduced.

  • How can we respond to client at scale & take action internally?
  • Aggregating data by (theme) to drive one metric
    • To be able to focus on one number and get the whole company behind is very powerful
  • Aligning between internal and external expectations
    • There’s difference between what people want/need
  • How do you scale that feedback for meaningful insights
  • Using the feedback to take action
  • Prioritisation of the feedback
    • Lots of feedback that we can’t action
  • How do you balance user insight/research vs feedback
  • Understanding basic level structures in place
    • Customer support/success are sitting on a goldmine right now
    • Weighing different types of feedback, e.g. Trustpilot feedback vs NPS which is easier; how do you prioritise that kind of thing
    • From a statistical point of view - what feedback is more important?
    • Everything you do needs to be fit within a commercial context
    • How do you validate the feedback: how do you know that you have a perceptional problem vs a real problem? How do you validate at scale?

  • Onfido: If it’s a detractor/promoter, the system will create a thing on Asana
    • Several thousand people. Only that many will respond, and only that many will agree to a call
    • Send twice a year
    • Allocate resources to whatever the key trends are
    • Try to solve the problem
    • Plot the trends on a bar chart; each bar is an item and it quantifies NPS impact.
    • Weighting is really interesting
    • Would like to link the customer NPS scores with financial metrics (e.g. LTV, acquisition cost, etc.)
    • Would be great to say: by improving NPS, we managed to impact LTV, acquisition cost, etc. - unable to do this yet
    • NPS: improving customer loyalty but it doesn’t necessarily move other metrics
  • If analysing manually, dangerous to have bias
    • Some outsourcing companies have automated sentiment analysis that humans review afterwards
  • You don’t need to use the same attributes but you need the same definitions for use across the teams
  • Biggest challenge is having everyone on the same page working together
  • Maybe segmentation should flex depending on the company objectives?
  • Important to have one person who is responsible for driving the internal feedback dissemination
  • Vision/user research/feedback mix for building the product feedback
    • You cannot have one without the other
    • Fake door experiment - let’s see how many people use it / do it (giving a sense for demand)
    • You should never be building things based on faith for too long
    • Tech is what enables operational teams to scale; how do we support growth vs refactoring internal stuff? How do we balance the two out?
    • Without operational excellence we cannot scale business
  • “Power of moments” by Heath brothers book - on bringing delight, what makes the peaks/troths
    • Focus on I like you to I love you, vs I hate you to I like you

  • Is there a clear vision?
  • Are we good at gathering feedback?
  • How are we prioritising? Weighting the data?
  • Is user research or feedback more important?
  • Qualitative vs quantitative data
  • Qualtrix (?) > own Delight(ed)?

Segmentation is not fixed. It will have to flex depending on where the company is.

It’s a journey. It won’t be perfect from the get go.

What is reassuring that no one has the perfect answer :) Good to know that we’re not missing a trick.

Everyone has that challenge of scaling it up.