Latest / The Retail Podcast / How to Embrace Real-time Feedback

Transcript

How to Embrace Real-time Feedback · The Retail Podcast

0:00
Hello and welcome to the Retail Podcast.
0:02
Today we're joined by Tim Waterton from Happy or not.
0:05
Rather than keep him hanging on the stream, let me bring Tim
0:09
onto the stream saying welcome to the podcast.
0:12
I'd be so surprised if someone hasn't seen your product in a
0:16
retail airport. You name it, your your product
0:20
and service has been there. But for those who don't know,
0:23
you just give us an overview what what Happy or not is all
0:26
about. Yeah, you're right.
0:29
We do tend to get seen typically in airports.
0:32
London Heathrow is the classic one for me.
0:35
We've deployed I think probably across 130 plus countries now
0:40
4000 brands. A few of the retail brands I
0:43
guess that would pop out would be Little Spa, Inditex, Zara.
0:48
So lots of varieties of bread, yeah.
0:51
But essentially people will have seen smiley terminals and
0:56
they'll understand how that kind of works.
0:58
But the the real principle behind what we do is very light
1:03
in the moment, richness, capture or feedback at scale, and
1:10
probably shouldn't be confused, if you like, by the simplicity
1:14
of the interface, I mean the Smileys are very engaging,
1:17
culturally agnostic, They're language agnostic.
1:20
That's kind of why they work that way.
1:23
But there are follow up questions behind that on
1:25
different ranges of our devices and we're able to get that data
1:30
back to the core in real time, slice it and dice set, match it
1:35
with other stuff and come up with a whole bunch of insights.
1:39
So the subtlety and the magic really happens behind the scenes
1:43
rather than a point of capture. But obviously there's really
1:47
quite a lot to making sure that we stand out as a brand and pop
1:51
out, if you like, from the background and generate the
1:54
volumes of feedback that give us the data to cure.
1:58
So I mean, look, let's let's let's just unpack that cause I
2:01
think there's a you, you, you, there's a lot of competition out
2:04
there in this space and and people take a different approach
2:08
to to feedback. So if I'm one of your, I mean
2:12
obviously the use cases will vary whether you're a grocer and
2:16
where, where you put the terminals and what you're hoping
2:19
to get. But what?
2:20
What do you find? What's the number one reason
2:23
customers come to you and choose yourselves?
2:26
What? What what?
2:27
What's their thinking and what just to add to that what did
2:31
they not think that sort of surprised them once you sort of
2:34
they got to know you a little bit better or they started
2:36
deploying the because everyone has the perception of I'm going
2:39
to get immediate feedback for example at point of sale if
2:43
someone had a good experience through their path to purchase,
2:47
right. I assume that that's that's top
2:49
of mind for a lot of folks, but what are the things that they
2:53
don't expect that they get from you?
2:57
That's a very good question. I think that there's so many
3:03
examples of deployment even within retail, because retail in
3:06
itself is so multifaceted, right?
3:08
So, yeah, you know it's high value electronics or whether
3:13
it's kind of very fast flow convenience store, fundamentally
3:18
different businesses for sure. So one thing is the real time
3:23
nature, so real time nature. And I've talked about Heathrow
3:27
in many situations, in convenience stores, in league
3:30
dietas, etcetera. Some about the voices actually
3:34
popped into the washing, right, Because nobody likes about
3:38
washing. That's a fact of life.
3:40
A lot of that that was employed in washroom, immediate, real
3:44
time feedback allows people to get in there and deal with any
3:46
situations they've actually need to deal with.
3:50
Terminals be real lot like live. As in if you are hitting you're
3:54
getting negative sentiment. Is it feedback live or is it a
3:59
collated? That I there's there is a
4:02
latency involved obviously, but we're talking like a couple of
4:06
minutes rather than talking about hourly chunk of data,
4:11
right. So it's not real time in the way
4:13
of kind of low latency trading on Wall Street.
4:16
We're not at that level. We're not going to invest at
4:18
that level. But it's it's meaningfully real
4:21
time in the context of an operational team on ground.
4:25
So things that people are surprised about, they're
4:29
surprised about when they're busy hours are the surprised at
4:34
the difference in perception of service from hour to hour.
4:41
I mean good examples attention example actually I think sees
4:45
them at Spar actually had an interesting experience with us.
4:49
She I familiar with Spa, right. Obviously I I grace a some
4:54
significant scale. First example that they found
4:58
out was, ironically, a broken fridge.
5:01
Right. So what actually happened was
5:04
they started to get a cluster of negative feedback as people were
5:10
leaving the store having come back to return mill that was off
5:16
ohh wow. And they were able to look at
5:18
the open feedback and sit there and realise that there was a
5:21
spike of people complaining about the milk.
5:24
They went down that aisle, checked the fridges and found
5:27
that one of the fridges would break and actually the milk in
5:30
that refrigeration unit was warm.
5:34
So guess you could find that any bunch of ways, but they got to
5:37
it quickly. Couple of things obviously you
5:41
know fridge back, yeah but apart from getting the unit fixed,
5:45
they changed their operational procedures to make sure that
5:47
they did a walk around and check the bridges on a regular basis.
5:51
Similarly, I I think again I've Susan's examples are always
5:55
good. The the other one was a cluster
5:59
of negative feedbacks simply because a member of staff was
6:04
particularly slow on the checkout.
6:07
Now that happens. It was a junior member of staff
6:11
so she was able to respond straight away by flipping that
6:14
person off the checkout, putting the more experienced staff
6:17
member on the checkout and then applying some kind of training
6:20
in the background to get that member of staff up to speed.
6:25
You know it's it's many embarrassed.
6:27
I think some organisation was a response actually over busy
6:30
periods, very, very busy period and allocation of labour is
6:35
often an issue, right? So whether we can get more folks
6:38
onto the tools, onto the checkout process when we need
6:40
to. But actually the other thing
6:42
that she did, which is now quite commonplace, I noticed, but she
6:45
she gave them headsets so they could communicate so they could
6:48
discreetly notify each other when they were getting overrun
6:51
and they needed help. How do people find this?
6:55
So sorry, I missed the last point.
6:56
Identifying pinch points? Really.
6:59
Yeah, so how do you what's the gap between negative sentiment
7:03
and attribute to negative sentiment?
7:05
So if you're picking up negative sentiment like at the checkout,
7:10
how were they able to what what what was the?
7:14
How do they realise it was actually the person checking
7:16
them out as opposed to, I don't know, the size of the queue or
7:20
something? Yeah, sure.
7:23
So it has a little bit to do with the variety of devices.
7:26
So we have different devices in different form factors.
7:29
So from the very lightest about devices which is a mini device
7:33
which is purely touch buttons and they feedback, it's not a
7:36
tablet based device that can be deployed individually on
7:40
checkout lines. So we have an example with the
7:42
supermarket in New York where they have 27 of their stores
7:47
breaked out without your not equipment and actually they have
7:50
a mini device on every single checkout lane.
7:52
So they can find out if they're getting negative sentiment on,
7:56
you know nine or eight or seven. They also have those mini
8:01
devices or all of the specialty food stands, so the fish stand,
8:05
you know, the cold meat stand etcetera.
8:07
So they understand if they have a problem there and they then
8:11
interestingly they complement that by using our tablet device
8:16
back of house in the breakout areas supported by QR codes on
8:20
the walls and on the lockers to get employees experience because
8:24
they understand that sink between employee experience and
8:28
customer experience and there's a very strong correlation slight
8:32
statistically and not just intuitively.
8:35
So it's a smart way for them to do that.
8:38
It's a smart way to give their employees voice.
8:40
And I think we probably all acknowledge that you know you if
8:46
you've got a happy customer, if you've got a happy employee,
8:49
that's no guarantee that you're going to deliver a really happy
8:52
customer experience. But if you've got an unhappy
8:54
employee, it's pretty much certain that it's not going to
8:57
be a great customer experience. How does that so when when you
9:01
look at experience and its relationship to conversion,
9:05
upsell and you know the financial side of things, again
9:11
how, what does that look like? Is there is there proven data to
9:15
say X amount of happy customers led to higher conversion or
9:21
repeat purchases? Yeah.
9:23
Coming from a CR, a position and a sales background, I wish there
9:27
was an absolutely simple formula to translate those to you
9:30
because it would just make it so much easier to explain to people
9:34
that genuinely campus and numbers behind it for you.
9:36
So one of our customers are actually a large convenience
9:41
store chain in the US Yeah, they're actually HQ in
9:44
California. They've got 500 locations and
9:47
then serving the across Washington, Oregon, Nevada, I
9:50
think Colorado as well. They have been running a project
9:55
with us for a period of time and multi phased rollout project
9:58
across their fleet, across the brand and like most customers
10:04
they wanna know what the impact is.
10:05
What are you gonna do to top line?
10:07
What are you gonna do to my bottom line?
10:09
And they've run a comparison test between a control set of
10:13
stores of exactly the same mix of configuration, square
10:18
footage, etc. And what they noticed was that
10:23
happy or not stores experienced between a 3% and 5% increase in
10:30
sales versus the control group. Ohh wow.
10:34
So obviously that would be specific to their business.
10:39
It doesn't always that's why I'd love that to be a formula across
10:41
the piece, but obviously it didn't.
10:43
But it is specific to their business and and I would
10:45
compliment them as there are there are company who look at
10:49
the data and therefore take action on the data.
10:52
So the data itself is worth nothing.
10:55
It's only when customers actually look at the data and
10:59
decide actually, yeah, we need to do some more training here or
11:02
we need to adjust our staff rostering here, etcetera.
11:05
But 3 to 5% is hard to buy in today's markets.
11:09
So obviously we're accelerating that project with them and look
11:12
forward to that being a great success.
11:14
Yeah, I mean, look, everyone's trying to do more with less,
11:17
right? Budgets are constrained,
11:18
inflations coming down but still high and and there's a,
11:23
especially in the UK, there's a cost of living crisis and in
11:26
some various other countries around the world.
11:30
So do you find that people naturally drift by wanting to
11:34
make the invisible visible? That's the primary sort of
11:39
engagement there, right? Because they're not.
11:41
What? What other metric or how else
11:43
are they supposed to understand how well that store estate is
11:47
performing? Yeah, 100%.
11:50
It's trying to surface that magical bit of data that allows
11:55
me to, I guess, make small incremental operational
11:59
improvements. They're going to net out at an
12:02
improvement on the bottom line. So that's definitely what
12:05
they're after. And he used visible and
12:07
invisible. You're right, it is invisible.
12:11
This crazily invisible. But the kind of feedback you get
12:14
from this stuff is really like having your customers as your
12:19
McKinsey giving you advice about what you should do with your
12:23
business. And they're not all right, For
12:26
sure they're not all right. But if you look at that in the
12:29
round and it's got statistical validity, I would definitely be
12:32
listening to them all day long. And by the way, it surprises.
12:37
There's a lot more positive feedback than there is negative
12:40
feedback. And you asked earlier about what
12:41
surprises do people get. That's actually probably the
12:43
biggest one. No one.
12:45
But they they they're desperately coming to the table
12:47
looking to how we're going to tune the business and what can
12:50
we do and what are we doing wrong.
12:51
Yeah. Only enough they find out an
12:53
awful lot about what they're doing, right because human
12:56
beings are generous souls that are.
12:58
I like to think so and and they share positive feedback probably
13:01
more often than they do negative feedback and are really smart
13:04
customers bake that in and make sure they read that back to
13:08
frontline staff from regional management and store management
13:11
and some of them wrap incentive programmes around it.
13:14
I would say the one thing not to do is make it part of a bonus
13:17
plan. I don't feed somebody with this
13:19
stick. It's not there to beat anybody
13:21
with. It's far better to turn around
13:23
and provide positive reinforcement than than any
13:26
negative. And how do you see in terms of
13:29
when you look at the evolution of how old is the company?
13:34
I companies on over 10 years old, now I'm.
13:38
Coming still young right in the in the context of things it's
13:40
still a young concept but I mean just from experience I know
13:44
building feedback loops has been that sort of magic Nirvana that
13:48
reads out like the one that you said about the bathrooms and
13:50
making sure you know bathrooms and grocers are you can walk in
13:55
there and and and and it's an OK experience.
13:58
I'm just curious what what you see the future, how you see the
14:02
future then in terms of what, what's the type of of a I was
14:06
the type of world we're moving towards in the feedback space?
14:10
Yeah. I guess there's a couple of
14:12
things in there. The first one I would pick out
14:15
is of devices originally. If you go back to that inception
14:21
point when Haiti walked into a particular store and repeatedly
14:25
got bad service and suddenly decided he was going to go and
14:27
build a company out of it, which I have to take it, it's pretty
14:31
impressive response, right? Yeah, but we obviously started
14:35
from purely button based feedback without having follow
14:39
up questions and having open feedback or collection of
14:41
contact details or all of that kind of capability, but that
14:44
they're all similar kind of things, right?
14:47
They've got us to the point where we understand what people
14:50
are thinking, why they thinking where and when.
14:53
So those four's are kind of covered were recently good
14:57
example of AI is a demographics capability.
15:01
So we lay it on that, at least a subtle nuance of The Who.
15:06
So basically by using the camera and the tablet, not storing
15:13
anything that we shouldn't be storing, but basically doing
15:16
facial recognition and being able to establish a demographic
15:20
fit, we were able to classify that data by age group ranges
15:27
and gender as well. Which in retail, you know what
15:30
it means, right? If I go into Abercrombie and
15:33
Fitch, and I think it's a dark place with loud music, frankly,
15:36
they're not that bothered because I'm not so much part of
15:38
their target demographic. Yeah, but they really do care,
15:42
right? If they 18 year old is in there
15:46
and turning around and saying this place is decidedly not,
15:49
yeah, so you know demographics is everything, it's it's all the
15:54
way through retail in terms of looking at football and video
15:57
analytics etcetera. So that's really off side of
16:00
your business in terms of what you're talking about there as in
16:04
the software side as in there's some form of display you
16:08
registered, you know the demographic details or is that
16:13
computer vision assumption of demographic details?
16:16
It's computer vision assumption of demographic details at the
16:20
point of feedback collection from variety of devices retailer
16:24
chooses to enable it etcetera. So it's another way, another
16:27
rich dimension. It puts The Who, yeah, alongside
16:30
the other 4 W And then the way I see it going is I see a whole
16:36
load of AI generated insight rather than pointing and
16:43
clicking around the charts and the analysis that's there.
16:47
Yeah, so we're pretty much a notification based society,
16:53
right? So what we really want to do is
16:55
receive a notification. We do that right now with things
16:58
like real time alerts, but I would like perhaps to subscribe
17:02
to the interest over customer service or store presentation or
17:09
anything about cleanliness. So I could subscribe to an
17:12
interest area for a range of experience points across a
17:17
number of stores, etc. And I could receive a
17:20
consolidated set of insight from that set of subscribed interest.
17:26
I can always dig back into the data afterwards, but I think
17:29
that's the way it's going to get because our attention spans are
17:32
so much shorter. Yeah, well now we're helping out
17:37
loads of businesses by providing A managed service.
17:40
So that's not unusual for SaaS companies.
17:43
We're deeply familiar with the data that we produce and we also
17:46
understand across different environments.
17:48
So lots of people are turning around and saying actually, can
17:51
you help us run service for us, Can you analyse the data, Can
17:55
you help us manage some surveys? Actually, can you do specific
17:58
education for our start as well. So we have a measure, analyse,
18:01
educate programme as a managed service which is there to quick
18:04
start, accelerate or do the happy little thing if you want
18:07
to look at it that way. OK.
18:09
And then so in in terms of so you you just see more AI LED
18:14
analytics on on basically what's happening in your in your store.
18:18
So rather than the the store manager.
18:20
But there's one thing I always think poor store manager
18:23
especially in grocery they get so many reports, right.
18:26
You get like a report on every single aspect of the store and
18:29
it's like what do I act on, right.
18:31
What's priority for me or for, yeah, for them in that in that
18:37
moment. And and I I guess that at some
18:40
point the AI will say, hey this is real priority because it has
18:43
an impact on bottom line which is what the in my mind that's
18:47
what feedback really should be able to, to help you accelerate.
18:51
So you're getting to that. Is this having an impact on
18:54
people checking out as much, for example?
18:57
100% and and the data that we produce.
19:00
Incredibly powerful in its own way.
19:03
More powerful when it's mashed together with a few other
19:06
things. So smart environment.
19:09
We look at football data as a data source.
19:12
We look at the labour management package to look at exactly how
19:17
many employee hours are deployed on the floor at any one point in
19:20
time. And we also take the sales data
19:24
and we then doing that we can generate kind of four subsidiary
19:28
metrics that are very, very insightful because kind of
19:33
counterintuitive actually in a lot of situations you look at
19:36
the data and you think this store is understaffed or this
19:39
store they actually really one being understaffed.
19:43
And in some situations you then compare it with another store
19:46
within the control group and you realise that actually another
19:50
store with less staff is serving a higher degree of football and
19:56
getting a higher degree of customer experience.
19:59
Interesting and. Actually speaking to somebody
20:02
very insightful comment from one of the guys at at the IGA and
20:07
the way he looks at it and he deals with this all the time, he
20:10
said actually management teams probably know that that
20:13
particular store is a problem. They probably know that it's low
20:16
engagement, they know it's underperforming, but they don't
20:19
have the data to point to and turn around and say why and what
20:22
should we do over it. So it's it's really about
20:27
providing the data that allows somebody to maybe confirm some
20:31
of the stuff that's in there. Intuitively, we're making it a
20:33
little bit more of a data-driven exercise.
20:36
So Tim, having worked with a lot of store managers and retailers
20:41
in general and obviously head office workers have a
20:44
perspective of the reports that they produce and you speak to
20:48
the store manager. But there's this thing about
20:50
data and action and the discrepancy between the two.
20:54
And obviously that's why I was really interested in how live
20:59
sentiment analysis potentially upscales your data.
21:06
I don't know if you have any thoughts.
21:07
Is that true or is that just or or not?
21:11
Are people able to get actionable on feedback?
21:15
Specifically, actionable insights that are data-driven
21:20
the act in the now. Yes, totally.
21:25
I think it's a really key point. Actions, the word that they
21:29
excite people, right, That's that's the best, that's what
21:32
it's all built on. And I think the fact that the
21:36
feedback is captured in the moment is the key to this one.
21:40
So I guess my example would be, we're all customers, we all go
21:45
into retail outlets all the time.
21:47
We all know what it's like when something's not working out and
21:49
we can see a resolution for it, but there's no action being
21:52
taken and that's up frustration. I can kind of measure it.
21:55
My blood pressure probably goes up a couple of pips and and my
21:58
pulse rate increases as well. So I could be looked up to
22:01
something that would detect it. But if I'm given the option at
22:04
that point in time to share some feedback, not irrational
22:07
feedback, but actually provide to answer a question, I think
22:10
this is the problem and this is what I would do about it.
22:13
That is a really valuable piece of business information in real
22:17
time. Now reality is I get through
22:20
that problem area, I get through that friction point, and I
22:23
guarantee that I'm not even halfway across the parking lot
22:28
and that valuable piece of business information is
22:30
evaporated. It's into the ether.
22:31
We will never ever get it back because the negative experience
22:36
has kind of launched somewhere in my head and I don't know
22:38
where that place is. Like, I'm not neuroscientist,
22:41
but it persists. It affects my behaviour in the
22:45
future, it affects my probability of coming back and
22:47
using the same store again. But I want the next in my life,
22:50
right? Whatever the heck that is.
22:52
But that piece of business information has effectively been
22:55
lost. Unless you captured it in the
22:57
moment. Unless you got it back to the
22:59
core. Unless you gave the operational
23:01
teams chance to react to it and do something good.
23:04
So 100% concur, It's all about the action.
23:08
It's not about the pretty charts and bits and pieces, it's all
23:10
about the action and I guess. That's where you would consider,
23:14
that's where you differentiate yourselves I would say, right,
23:17
yes. Totally.
23:18
I've worked. I I think I mentioned at the
23:21
beginning this I'm a data guy, but over data now time it's been
23:23
my background. Yeah, so work with huge
23:26
warehouses, marks, lakes, whichever ones you want to call
23:29
them these days because whichever turns to become more
23:31
fashionable. But you know a huge terabytes of
23:34
data sliced every which way from Sunday, millions of charts of
23:37
different formats, etc. The reality is sometimes people
23:41
ask the question, why is the title in green, that they're
23:45
really not focused on it. We're we've got far too much of
23:48
that. Yeah.
23:49
So yeah, platforms that just serve up a huge amount of charts
23:53
I think could possibly less interesting to retailers like
23:56
you said, they they get a lot of reports.
23:58
A heck of a lot of reports, yeah.
23:59
Absolutely got a lot. On their shoulders.
24:01
So yeah, it's actionable insight.
24:02
That's the key. OK, come to the end of the
24:08
podcast where can people learn more about, you know, feedback
24:11
loops and and what you guys do and help you.
24:14
Are you an RF? Are you at any of the major
24:16
events coming up? We have.
24:18
Been to NR. If we tend to not do a whole
24:23
bunch of events, we get a lot of people coming to us, actually.
24:28
OK. And I think if you've noticed
24:30
this, particularly in the retail space as well, references
24:33
everything. Yeah, People trust what they
24:36
hear from other folks. And we get a huge amount of our
24:41
inbound business from people having positive experience with
24:44
us and sharing that with other people.
24:45
So I would encourage anybody to literally hit the website
24:50
havefearnot.com, take a look at it, tell us what we can tell you
24:55
more about what we do and have a conversation with us because
24:59
we're actually about trying to help and prescribe the right
25:04
solution for people. We're not here to sell you
25:07
something that doesn't fit within the business.
25:10
If we're not going to make an impact, if we're not going to
25:11
move the needle, then we're, we're not going to do business.
25:15
But if we can, we've got Tim. As you can see, I can carry on
25:18
asking questions forever, but I really appreciate you taking
25:21
some time out of your day to come and have the conversation.
25:25
Absolute pleasure, Alex. Really enjoyed it.