How to Set Up an Analytics Team | Analytics Organizational Structure | @ValueDrivenAnalytics
1 Video View·Dec 1, 2024
Building a high performing analytics team is not easy, but these tips will help set you up for success.
Are you looking to build an analytics team, but have no idea where to start?
Hi, I'm Brian Amrein, co-founder and principal analytics consultant at Value Driven Analytics,
and today I'm talking to you about how to build a high-performing analytics team.
Now the first thing you need to do when you're building an analytics team
is determine what your goals are.
And what your goals are could influence how you set up that analytics team.
For instance, you need to determine whether you're primarily looking to build reporting for stakeholders
so that they can see trends in the data,
or whether you're looking to build more of an ad hoc analysis team
that can look into business questions for your stakeholders,
dive into what impact different initiatives are having and why you're seeing certain business trends,
or whether you're looking to do some predictive or prescriptive analysis
to really do some more advanced analytics using data science techniques.
Now maybe the answer is all three of those.
I'd like to have a team that builds interactive dashboards for my stakeholders to drill into.
I'd like to have a team that can look into ad hoc business questions.
And I'd like to have a team that can build predictive and prescriptive models.
Well now that you've answered that question,
and all three of those are high-value things that your business probably should be investing in,
it's time to figure out what the optimal setup is.
Now there's a few different setups.
Some organizations decide to have a dedicated reporting team,
and a dedicated ad hoc analysis team,
and a dedicated data science team.
The development initiatives and the topics may look different depending on those areas,
and so why not structure teams around those?
Now other organizations take analysts,
and they actually give them a mix of reporting projects and analysis projects,
and even sometimes data science projects,
as they learn to perform more and more advanced analytics techniques.
Now what's neat about that is it gives the analyst a variety of things in their area,
and if they're assigned to a specific stakeholder,
they can handle everything from reporting to analysis to data science.
Perhaps if you're a larger organization,
and you're looking to build out a team of maybe 20 or more analysts,
perhaps you have that level of specialization where you can have a sales reporting analyst,
and a sales ad hoc analyst,
and a sales data scientist.
But if you're a smaller organization,
it very well may make sense to have an analyst dedicated to the sales team, for instance,
and they work on the reporting, the ad hoc analysis, the data science,
anything that comes up in that area.
Now what's neat about that model too, that combination model,
where an analyst works on multiple things,
is it doesn't limit them.
For instance, and a lot of analysts do aspire to eventually learn data science tools.
There's no barrier there.
That's the work for the data scientist.
Now there's ways to facilitate development,
even if you do have a more segmented approach.
Maybe you pair an analyst with a data scientist
if they want to become a data scientist eventually.
There's some mentorship there,
and eventually when they're ready,
they formally switch over,
and there can be ad hoc projects.
So you can make that development piece work either way.
But I do find that a lot of analysts like the variety of working on dashboards,
and analysis, and model building.
In fact, some projects actually require two or three of those techniques.
For instance, maybe there's a project that involves building a model.
In order to implement the model,
an interactive dashboard needs to be built
to highlight what the scores are coming out of the data science model.
And then on top of that,
there might need to be an ad hoc analysis
to assess the benefit that you're actually getting
from using the model versus the old way of doing things.
So in that example, all three of those are necessary,
and so it's nice if you have an analyst that has all three of those skills.
Of course, there can be collaboration across analysts as well,
where they divide and conquer each of those aspects.
There's no right or wrong way,
but those are some considerations to have.
Even if an analyst does have data science skills,
some will certainly say,
yes, I want to do all data science, 100% data science now.
But others will actually still say,
well, I like data science.
I like that I get to use cutting-edge techniques
to automatically detect trends in the data,
but I also enjoy creating interactive dashboards.
There's the artistic aspect
and making things aesthetically pleasing that I really enjoy.
And then on top of that,
sometimes I like to perform ad hoc analysis
because of the quick wins that we can get with things like that.
And so some analysts will actually, in the long term even,
enjoy working on a variety of projects.
It's important to get to know your analyst,
and maybe you're able to facilitate development
and facilitate career interest
by having some analysts that want to focus on one area do just that,
and other analysts that prefer to span their work across those areas
do that as well.
You might also want to think about how much reporting you're going to need
versus how much data science modeling you're going to need
versus how much ad hoc analysis you need,
and that'll determine what kind of roles you need to hire,
especially if you're going to go with the very segmented approach
of having data scientists focus on data science models
versus the analyst who is kind of a jack-of-all-trades,
does some dashboarding, does some ad hoc analysis,
and does some data science.
Now, if you go with that jack-of-all-trades model,
you're probably looking at at least an analyst role,
which tends to be, sometimes this is used as an entry-level role,
but it can also be a mid-level role.
Sometimes people, organizations will sometimes use analysts too as well.
But more likely, if you're going to go with someone
who has a variety of skills,
dashboarding, ad hoc analysis, and data science,
you're probably looking at a senior analyst
who would have all of those skills,
or you could even give them the title of data scientist
with the understanding that they'll also be doing
some ad hoc analysis and dashboarding.
So that's maybe a good route to go
if you are going to be looking for a team of jack-of-all-trades type of analysts.
Now, if you are going to go with a specific team that does reporting,
you would probably call that role a reporting analyst role,
or you could just call it an analyst role as well.
Now, most of the time,
probably the most common title you'll see is analyst or data analyst,
and reporting analyst would work if they're focused on reporting.
But if you're looking for someone more senior,
you might go with an analyst too,
or senior analyst tends to be a popular title as well,
senior reporting analyst.
And if you're looking for someone more junior,
you might create a junior reporting analyst role.
Let me recommend this, though,
when it comes to hiring junior analysts
or even analysts in general with less experience.
If you're a very small organization
and maybe you're just planning on hiring one or two analysts,
I would highly recommend hiring some analysts with some experience
because if you are going to go with a more entry-level analyst position
like a junior analyst role,
you're going to need somebody to mentor that person
and validate their work.
You would not want to assign a junior analyst
to a manager who doesn't have much analytics expertise.
You're just asking for trouble.
They might be at a point where they're still learning,
and if you don't have someone who can robustly assess
their methodology and the data accuracy
and also develop that analyst,
one, they're probably going to be more likely to leave
to go to a place where they can be developed more,
and two, you may be making decisions
based on a bad data analysis setup or inaccurate data
and not even know it.
So it's really important that if you do go
into that junior analyst pool,
which has a lot of benefits, right?
They can grow with you.
You can teach them from the ground up
and probably less expensive than a more senior analyst,
lots of benefits,
but you need to have the resources to train them,
so look out for that.
If you're a small organization,
you're just hiring a couple of analysts,
hire more senior analysts.
Now, if you're looking for someone
to perform ad hoc analysis,
get titles like junior analyst, analyst, analyst two,
senior analyst,
depending on their years of experience in the industry,
those are going to be good titles,
and if I had to put some numbers to it,
I would say a junior analyst,
probably someone coming right out of school
or new to an analytics position,
and I would say for an analyst,
it's probably someone sometimes also right out of school,
but up to maybe two, three, four years and beyond.
There doesn't have to necessarily be a limit to it,
but then when you're thinking about analyst two,
you're probably looking for
at least one to two years of experience.
Senior analyst, probably two, three, four,
or more years of experience.
Everyone's on a different trajectory.
Really, what I would recommend basing a title on
is where they're at from a skill perspective,
and there's one way that I would highly recommend
assessing where someone's at
from a skill perspective in an interview,
and that's using case questions.
Whether you're looking for a reporting analyst
or an ad hoc analyst or data scientist,
give them a true business challenge
that you would expect them to work on
and see how they would approach it
with the analytics tools that you expect them to have.
Now, if you maybe don't have
a lot of analytics resources on your team right now,
you don't have the analytics expertise
to come up with that question,
or if they give you an answer,
assess whether it's a robust answer or not.
Value-driven analytics can help with that.
We offer one-hour analytics consultations
that can be used to assess potential candidates.
If your organization would like to do
some advanced analytics,
you're going to need a data scientist.
The most popular title for a data scientist
is simply data scientist,
but you will occasionally see organizations
that'll use a role like junior data scientist.
If there is someone who has gone
through a data science program,
whether it's through a university or boot camp
or has some previous experience in data science
but maybe not a whole lot,
you may consider a junior data scientist
a title for them until they can prove
that they have rigorous skills in that area.
Likewise, if you have someone
who's been in the data science industry for a while,
perhaps you would give them
a senior data scientist role.
And if you have someone who's really performing
at a level where they're essentially
a leader in the industry,
coming up with perhaps new techniques
and maybe even algorithms themselves,
you might even consider the title
principal data scientist.
Especially if you're planning on hiring a large team,
you're going to also want to think about
what your management structure could look like.
What I would highly recommend,
I have an entire video around this,
is creating an analytics center of excellence,
but within that analytics center of excellence
you may need to have specific teams.
You may have a team of sales analysts
and marketing analysts, operations analysts,
supply chain analysts,
all the different divisions the team
is expected to support.
You could organize it based on functional area,
I would recommend that,
but it's also very common for organizations
to structure their analytics center of excellence
by having a dedicated reporting team
that works on reporting,
analysis team that works on ad hoc,
business analysis questions,
and a data science team that focuses
on the data science modeling.
That's not a bad structure either.
But whatever it is,
you're going to need potentially leaders
for those areas,
unless you only have four or five analysts
that would probably report up into a,
could be a manager or senior manager or director.
If you do have this analytics center of excellence
for you,
either based on what they're focusing on
reporting, analysis, data science,
or the stakeholders they support,
sales, marketing, supply chain, for example.
Either way,
you're going to need some leaders on that team.
And so if you,
if they're only leading maybe one,
two, three analysts,
you might use a title like analytics team lead.
If they're managing more like three,
four, five, six people,
maybe you consider that to be a manager
or senior manager role.
If they have more experience with management,
you could even consider a director role
in that case.
And as analytics leaders grow,
sometimes you will see organizations
that have a senior director of analytics role
or even a vice president of analytics
or in some organizations
that are really data driven,
they'll actually have a chief analytics officer,
chief data officer that report up into the CEO.
If you're really looking to invest in analytics,
maybe consider some of those titles as well.
You've planned out what you want your analytics team
to look like.
And now it's time to perform
the very challenging step
of actually hiring your analytics team.
It's challenging,
especially in today's environment
where there's a lot of demand
for analytics talent.
But the very first thing you're going to want to do
is find an extremely,
extremely good analytics manager.
The reason this is so important
is it can make your job
as a leader of an analytics group
very challenging
or very easy
if you find a really good analytics manager.
And what you're looking for
is someone with a robust knowledge
of analytics, analytics tools,
analytics techniques.
Again, we recommend using case questions
to assess that.
And again, value driven analytics
can help with that.
But you also want somebody
who doesn't just keep that knowledge themselves,
but is passionate
about training others,
training their eventual team
in those skills as well.
Someone who can validate their work
and continue to develop them
and keep them engaged in that manner.
So it's important to find
a good analytics leader first.
Now eventually,
that analytics leader can help you
build out that team even further.
If especially it's someone
who has some really strong
analytics skills,
you may be able to hire
some analysts that are
coming in with very little
practical experience.
You could go to career fairs
at certain universities
and really target people
who have majored in a quantitative major.
Or if you're looking for
maybe more experienced analysts,
you just look for people
that are in a current analytics role,
look for some of the titles
that we just walked through
and target them that way.
Now if you're struggling
to find analytics talent,
like a lot of organizations are,
please reach out
to Value Driven Analytics.
We can help you with our network
to find analysts
with rigorous methodology
with really good analytics skills.
I'd like to walk through very briefly
some of the skills
you need to be looking for
when you're looking
to hire your analyst team.
The most important one,
the thing that is
very challenging to learn
or takes a lot of time to learn
is an analytical thought process.
You could look at it
as a general affinity for math,
someone who can take
a business challenge
and turn it into
an analytics challenge
in a robust way,
an analytics requirement.
That is such a hard thing to find
and if you can find that,
it's very likely
that some of the other skills
that you'll be looking for,
like coding
or finding experience
with using an interactive
dashboarding tool,
that that can be learned,
especially if you have
someone more senior
who's willing
and passionate
about training new analysts
in your organization.
So look for that
analytical thought process
perhaps more than any
of the other skills here.
It's also important
to have a solid attention to detail.
You don't want someone
who is not going to be
checking their work
or overlooks those aspects
or else you could be
in a position where you're
making a bad decision
based on an inaccurate analysis.
You also want someone
who has the ability
to communicate technical concepts
to non-technical stakeholders.
A lot of analysts
are responsible
for presenting their work
and oftentimes
to a non-technical stakeholder
and they need to be able
to do that without getting
in the weeds
to really be able to focus
on the insights
and even more importantly,
the actions coming
out of their analysis.
You do also want to look
for somebody
with some coding experience.
Ideally, again,
this could be learned
by someone who has
a good analytical thought process
but if they already
have some experience
with something like SQL,
Python, R, SAS,
or other coding languages
that are used
in the analytic space,
it's really helpful
because it can help
increase their productivity
compared to using
a tool like Excel
that tends to be more manual.
This is maybe surprising
to some folks
but analysts need
to be creative.
They're given
business challenges
and they are, again,
as we talked about
in the analytical
thought process skill,
asked to translate that
into an analytics approach
and that takes
some creativity,
especially when you're
designing dashboards.
That can take
some creativity too,
some almost artwork
to create a dashboard
that's aesthetically pleasing
but either way,
even if you're
a data scientist,
you need to be creative
in the way that you
create an analytics approach.
It's also nice,
maybe not needed
for every role,
but it's nice
if you've got someone
with some leadership skills,
even informal leadership skills,
the kind of analyst
who is willing to
proactively identify
and pursue analyses
that will help
your organization.
Wouldn't that be nice?
That's a great bonus.
You need someone
who's going to get
their work done on time,
who can manage
usually multiple projects
as analysts are often
expected to do
and going along with that,
you need someone
with some work ethic
who's going to work hard
to get what needs
to be done,
done on time
and with validation
along the way.
Getting all of that done
takes some work ethic.
Getting a little more specific
on the technical side,
I've put together a list
over here on the right side
that shows some of the skills
you might include
in your job description.
You definitely want
to put in there
that you're going
to be responsible
for doing some analytics
or data science
and whether they're
an analyst
or a reporting analyst
or a data scientist,
they're going to need
SQL skills,
especially if they're
a reporting analyst,
they'll need Power BI,
Tableau,
maybe Qlik,
just experience
with an interactive
visualization platform.
When it comes to
interactive visualization,
if they know one skill
like Power BI,
it'd be easy to learn
Tableau and vice versa.
You might say experience
with at least one
of these interactive
visualization tools,
Power BI,
Tableau,
Qlik, et cetera.
Same thing applies
when it comes to
advanced analytics tools.
You really want
to have these on there
if you're recruiting
for a data science role.
Python, R, SAS,
those especially
I would recommend.
Maybe even if you're
looking for a data analyst,
especially someone
who's going to be
performing ad hoc analysis,
perhaps you'd include
those as well.
For a reporting analyst,
it's not as critical
to have these skills
for advanced analytics
like Python, R, and SAS,
but it certainly
doesn't hurt,
especially if you're
looking for a data scientist,
you're going to want
to include keywords
like data science,
statistical analysis,
modeling, forecasting.
All of those are tasks
that a data scientist
would be responsible for.
And especially if you're
looking for someone
to perform ad hoc analysis,
maybe even a data scientist
as well,
you might want someone
with experience
with A-B testing.
This is one of the
most robust ways
to assess the impact
of an initiative,
which in some cases
could be the deployment
of a new model,
which is why you might
want to include that
even on a data scientist
job description.
After watching this video,
I hope you're more
motivated and more excited
to build out your analytics team
than ever before.
Remember,
it starts with
determining the goals
for your analytics team,
which then dictates
how many reporting analysts
you'll need
versus ad hoc analysts
versus data scientists.
Or again,
maybe you go with
a jack-of-all-trades
analyst style.
Whatever it is,
remember to put
a lot of focus
on finding
a high-performing
analytics manager.
That is so important.
Because then from there,
they can help you hire
the remaining analyst
on your team,
looking for characteristics
like an analytical
thought process,
coding skills,
good communication skills,
and more.
We talked about
some of the skills
you might put in
your job postings
and how it's very important
to use analytics
case questions
to truly assess
whether the analysts
that you're hiring
have the skills
to take a business challenge
and translate it
into an analytics requirement.
Now, I'll be honest,
building a high-performing
analytics team
is a challenge.
It takes hiring
the right analyst
and it takes
developing them,
keeping them engaged,
sometimes even
performance management.
It can be overwhelming.
Well, value-driven analytics
can help you.
They can help you
build your team,
recruit analysts,
train them,
and even manage the team.
So if you're looking for help
as you build out
your analytics team,
please check out
our website,
Value-Driven Analytics,
where you can learn more
about our analytics
team management services
and consider giving us a call.
I want to thank you
for watching today.
I hope you got
a lot out of it.
And if you did,
please consider
liking the video
and subscribing
if you'd like to see
more videos like this
in the future.
Now, if you have
any questions
about what we covered,
please consider
leaving a comment.
I'll try to respond to it
as soon as I can.
One last thing.
If you or your organization
ever need anything
analytics-related,
an analysis,
a dashboard build,
maybe a data science
model build,
perhaps deeper training
on this very topic
or other analytics topics,
even a complete
analytics transformation,
please consider
Value-Driven Analytics.
Check out our site
at
valuedrivenanalytics.com
and if it resonates with you,
give us a call.
We'd love to help you out.

