Executive summary
For any organization offering services to users, the Shift to Left model is
the roadmap to transform its way of designing, providing and supporting
services. Starting at a foundational organization of four levels of Support the
S2L model shows how important is for any organization to initiate that journey,
set priorities and objectives. Either if your company is a startup, has escaled
or is consolidated keep this model in mind to make your trip to a user
centric strategy in a faster and more confident way.
Introduction
With this title I’ve recently
participated speaking about our experience in the S2L approach at the Customer Service and Experiece summit organized
by Incite Group in London in September 2019.
Madrid Digital is a public company
which provides ITC Services to Comunidad de Madrid Regional Government and
Administration.
Comunidad de Madrid is the central
region of Spain, representing about 20% of the total Spanish GDP and having the
highest GDP per capita of the country. It’s population of 6,6 millions has
access to public services provided by a public administration governed by the
regional Government, to name a few Health, Social Care, Education, Justice,
Employment and so on. More information here.
Those public employees are the
users for Madrid Digital and in the last term all citizens benefit of our services. In this
context, Madrid Digital is directly providing services to more than 165.000
users and also in some aspects to potentially the whole population mentioned
above.
We support more than 80.000
computers, 45.000 printing and scanning devices and 87.000 fixed and mobile
telephone lines. We are supporting more than 1.200 information systems or
applications that our users use daily thorough the more than 3.500 sites where
the public services are provided. Finally we also support more than 9.000 network
nodes including switches and wifi access points for local communications.
S2L reference model
Most of organizations organize
their operations of support in what I’d refer as a traditional way ranging from
Level 1 to Level 3.
At level 1 we deal with
the reception of user cases or requests, their register, solution or
escalations, depending on complexity of each one.
At level 2 we deal with
remote or on-site cases resolution that can’t be solved at level 1. It’s
obvious that remote L2C is preferred over on-site L2P.
Finally, for specific issues or
most complex ones, we have level 3 that deals typically with IT
infrastructures and applications cases.
The S2L model bring us three
additional levels of “support”:
- Level 0 deals with
self-service or self-register capabilities for users, introducing web/mobile
channels for support. Additionally we put at this level the enhancement of our
IVR considering it a very real attention channel for our users. With this level
we improve our user readiness to self service avoiding the need to be assisted
by Level 1 staff.
- Level -1 is related to
monitoring and automation. Monitoring services from a user experience
perspective in first term and extending it to most critical infrastructures.
The use of auto healing mechanisms complements the readiness to be able to
quickly restore services when critical events arise.
- Level -2 includes SW
and Services development and engineering and deals with software quality
assurance and proactive patching of workplace devices for stability and
security purposes.
The objective is to set priorities
and action plans to migrate resolution of issues from higher levels of support
to lower layers of support, closer to the user. Level -1 or -2 have been
referred as “support” but really don’t solve issues but prevent them from
occurring in case of level -2 and help them to be solved faster in case of
level -1. At these levels we should identify which actions are we doing yet or
need to be done to we aware of its contribution to make the best User
Experience.
Starting point
At the beginning it’s important to visualize what is your initial picture.
In our case we were surprised to see a gap in these charts as we expected to
see that we were solving tickets remotely. After analyzing the data we found
out that we were using our ticket information system differently as we first
thought. We also introduced slight changes in our technical groups of support
procedures.
At this point we did not have any evidence of Layers 0 to -2.
One year later...
After one year of the launching of the web channel for user support we
acehieved a pretty good rate of web channel usage of 9% in average. We did
communication campaings to our users based con email messages and also adapted
our IVR messages to inform them.
Despite of having about 400 service indicators we still discovered new valuable ones to measure how efficiently we were solving user cases. This process is very interesting because every week you become more aware of interesting findings. For example, we discovered that depending of user cases there are types that are more efficiently solved than others and you can readjust the procedures of less efficient to the most eficient ones. This indicator was named ATI (Added Time Index).
Layer -1 deployments
Regarding Monitoring and automation, we started it from a user experience
perspective. Than means to focus on methods of monitoring that track the
avaliability of most critical services. This has been demonstrated to have the
highest benefit/cost ratio of monitoring just because when you have an issue at
this level of monitoring it has the highest probablity to be affecting end
users services.
Classical monitoring of infrastructures is much more difficult to be correlated with the impact of IT issues in services.
This can be clearly seen in our main monitoring screen that is our first level of monitoring.
When one of these spots is red it is because one critical service is down
or affected and we have to start inmediatly the service recovery procedures. We
also provided this view to our Level 1 team at the call center to be able to
give information to users when they phoned because a critical event. This was
highly important to remove users uncertainty when they reached our call center.
This source of monitoring is also used to inform users in our IVR messages to
inform them about major outages.
Then we focused in critical infrastructures monitoring as an important complement to quickly diagnose service incidents. When we confirm that a service is affected at our first level of monitoring, this second group of infrastructure metrics, help us to diagnose critical events much faster.
Other important set of actions at S2L level -1, were related to notification of events in real time. We had clear that to react inmediatly to critical events, they had to be notified inside our organization by any available means such phone, email, SMS, instant messaging notifications.
We finally completed this set of actions developing a mobile app to notify
major or relevant incidents. While the above methods were based on service and
infrastructures monitoring, this was based on our ticketing information systems
monitoring, selecting very carefully the most critical cases to be displayed to
get rid of the risk of spaming our support channels.
Regarding automation, to name only specific actions, we have implemented
the automatic opening of tickets from diferent sources of monitoring, self
restore of well known critical events and automatic phone calls to technical
staff specially important on non bussiness hours to restore services when most
users are away from the office, and have then available when they start their
job.
These automatic actions again have to be triggered by very carefully
selected events to maintain a high efficiency in our service recovery
procedures.
Level 0 deployment
As we have mentioned above, by the end of 2017 we achieved an average 9% of
web channel ussage. This was pretty different in education users. As it can be
seen in the picture, we were able to slightly increase it (from 1% to 4%)
through traditional communication campaings email based.
At the same time we launched new user centric services such Raices for
education users we announced them that the preferred chanel for support would
be the web. That´s how we achived an huge increase of use (from 4% to 27%) that
was also caused by the launching of the new application but it was enough
to give an average use of 7,6% by the end of 2017 for education users that
remained in a pretty acceptable rate of 12% the following 2018 and 2019.
Time and cost implications of S2L strategy
In the study for Madrid Digital regarding time we foung out that cases were
solved between faster and much faster times when done at level 1, precisely
between 20 and 80% faster for incidents and 15-80% for requests.
Regarding cost we found out similar results, giving cost that were
multiplied by an average of 2 when we raised the levels of support. That can be
clearly seen in the following picture.
Time and cost were important parameters for Madrid Digital while time was also an important one for our users.
User feedback
After time and cost study there was something important missing. We could
theoretically define our SLAs for support services that could be reached and
still have user complains and on the opposite hand have defaulted SLAs and an
acceptable user preception. So it was neccesary to ask our users about the
service.
We implemented a user feedback channel inserting in each email notifying the solution of cases to our users, a link to a simple web form regarding the main aspects of the service.
Thats how we obtained an average score of 3.8, 3.56 and 3.29 (in a scale up to 5) for L1-L2, Infrastuctures and Applications support groups respectively.
So from a time, cost and most important user feedback perspective we can clearly conclude that S2L model is te way to a User Centric Strategy.
Takeaways
To conclude our study we can say that:
1. You can have very remarkable insights if you apply this still theoretical S2L model to your organization to set you User Centric strategy. Doing so you will have a powerful tool to guide your support and non support actions.
2. It’s important to set your starting point and measure all time to track
your progress in your roadmap.
3. Start your monitoring from a user experience perspective complementing
it afterwards.
4. You can increase the adoption of new support channels when launching new
user centric services.
5. If you consider your organization user centric set a user feedback
channel and motivate a permanent increase and analysis of it.
6. Improve and evolve your model defining and measuring actions and
indicators related to L-2 through L3.
7. Set and track objectives and indicators at each stage of your strategy
deployment.
