Analytics

Depending on your perspective, you may or may not want to know how times you were presented with a set of analyses that could have had a meaningful impact on your business, but for one reason or another (which I’ll talk to in a moment), you didn’t realize it.

Most entrepreneurs, marketers, and business leaders whom I meet know that their metrics are crucial to their success.  In particular, they know how those metrics set benchmarks and then inform decisions for how to improve business performance.  But in talking to these same business leaders, I’ve found that most of them are frustrated by what their analytics team provides to them.

See if any of the following sound familiar from your meetings with your analytics team:

• A brutally painful spreadsheet, printed out at 50% of normal size
• A table that takes 15 minutes to understand
• Some cool graphs and data points that don’t have anything actionable associated with them, but they’re really cool, aren’t they?
• The analytics person delivering a report but having spent zero time thinking about *how* to take action on the data
• Reports that flat out just don’t make sense because no one spent the 5 minutes looking at the data to ask if it makes sense.  Have you ever seen site conversion jump from 5% to 76% in 2 days? Yea, neither have I.  But the report you looked at last week says it did…
• A disconnect between what you thought you asked for and what was delivered (this is certainly not exclusive to analytics)
• How about this scenario: at the end of a metrics meeting you find yourself saying, “Here, you just aren’t getting it.  Let me draw out exactly what I’m looking for.” Then the analyst sits there somewhat shaken, pretends like they understand, doesn’t asks questions, and leaves saying something like, “Ok. I see what you’re saying. I’ll get you something right away.”  (The other two likely scenarios are that you call in the CFO to help or that the meeting ends as another is about to start with something like, “we’ll pick this up later.” 2 weeks later, it still hasn’t.)

Let me be clear about something – this is not about taking sides.  The only side I’m on is that I want businesses to leverage their information and people as much as possible.  And as I’ll get into shortly, there are ways to make significant improvements to avoiding the above scenarios.

During my first few years at Beachbody, I helped to build the strategic analysis team.  Of course I’m biased, but I think that was one of the higher-performing analytics teams in the industry.  At the same time, a good amount of that list above comes from first-hand experience, typically something that I did myself.  The rest has come from helping clients build, leverage and develop their analytics people.  And as much as I’ve grown into a broader marketer, analytics still holds a special place in my heart.

In this article, I’m intentionally going to focus on the relationship and people side of working with your analytics team.  Obviously, there are a host of issues that are crucial to getting what you want – hiring, data integrity, tools, etc.  But in my experience too often the people side of working with analytics is ignored.

First off, to those of you who think you have a one-person analytics team, that’s, well, not accurate.  That person likely works with someone in IT and also pulls data out of vendor platforms.  They may be the one you go to, but it’s important to realize that analytics is never just one person nor one department.  For simplicity’s sake here, however, I’ll be using the term “analytics team” even if that means that one person for you.

I have an entirely separate post about hiring analytics people here, but even if your team comprises “what good analytics people look like” – good at Excel (or at least says they are), something technical in their college experience, and claim to have experience in statistical tools or Google Analytics – too many marketers end up wanting more from analytics.

Are we clear on what exactly Analytics is?

Whether it’s differentiating between buzz words like BI and Big Data, or doing so with some commonly-used terms like reporting and analytics, it’s important to make sure you and your teams are aligned on what these terms mean to your organization.  You have your understanding.  Likely people have worked elsewhere, and these terms have meant something else.

If this is already starting to feel like work, I’d say that’s entirely correct (I hear this often at this point in the conversation when I’m working with clients). And just like everything else in your business, making sure you are getting what you want from your analytics team takes at least a bit, if not a lot, of work.

At the core, however, many people don’t realize just how much is involved in building a high-performing team or just haven’t put the attention and follow-up in the right places.  Hopefully this article helps in those regards.

Now, on to the real matter at hand.  It’s a rare case where the analytics team is trained, in two specific areas.

First, there often is very little training about the business as a whole. Sure, you may have an onboarding process, and perhaps you’ve even talked about your company strategy at your all-hands meetings.  But I’m referring to a more granular view of the business – what historically have been the key levers, where are the known pain points, and so on.

Too often it feels like the relationship between business leader and analytics is solely transactional. Meaning, when you ask for something, you are likely literally asking for the specific thing you want. Or at least you think you are.  No real context is provided. Then when you get something back, it’s not what you really wanted.  And you’re wondering why you have to do all the thinking about what the report you just got means.

Your people will appreciate the time you’ve taken to explain the business.  More importantly, having broader context oftentimes results in their identifying opportunities simply because their eyes were open.  It may seem like a random comparison, but I liken this to how, once you’ve just gotten a new car, you start seeing them all over the place.  When in fact they were always there.  Now, you just had a reason to actually “see” them.  Bottom line, help your people know what to look out for.

Second, it’s a rare analytics team that gets focused training on how to deliver actionable insights in a way that someone else can understand. Because isn’t that ultimately the whole point of business metrics and analytics?

Fortunately, both of these two areas can improve.  With much more on that second point below.

So what’s the solution?

Below, I address the biggest areas of unmet needs I’ve seen in organizations.  Not surprisingly, there are a combination of factors that help to set up a great relationship between you and your analytics team.  One in which you’re getting the information and insights to help move the business forward.  And one where your teams are satisfied, highly-motivated, and high-performing.

Here’s what you can do from the business side 

1. Reporting vs. analytics

Reporting is a bunch of information put together and delivered to you, likely via email.  Analytics means taking that same information and providing real value.  It means insights and suggestions on what the data means.  This isn’t to say that your folks should necessarily come to you with new landing page designs or ways to optimize your call center scripting to improve customer satisfaction.  But it does mean to show up being able to describe what the information says and where there are areas of opportunity.  At a bare minimum, nothing you see should show up without a summary of key points.

2. Set the bar high.  

It doesn’t matter how your company used to operate or what the norm was at a prior employer.  Demand more of your people.  Tell your people to come with ideas, implications, and next steps.  They may not be high-ROI ideas at the outset, but just like any muscle, that skill has to be nurtured.  People need to be clear about what is expected of them.  And then hold them to that.

3. Remember, you’re typically dealing with introverts

One of the greatest challenges in effective communication between marketers and analysts stems from a difference in personality types.  Most analytics folks are introverts (whereas most marketers are the opposite), and so the analysts may take some time to process information or may feel less comfortable with conflict or challenging what you have to say. (Again, this is relevant for more than analytics people, but particularly so in my experience.)

They often struggle with knowing what the best means is to communicate findings – in person or email.  Sure, it’s on them to be proactive, but I can almost guarantee that if you ask them what they are working on today, there will likely be something that you are interested in.

4. What motivates them

Each person has their own irrational passions (more about that in another post).  Typically, analytics folks want to be able to show what they can do and want to know that their work lead to something impactful.  They want to feel like they are a part of the team, not just some back-office Excel or data monkey (perhaps one of the pejorative terms I’ve heard them referred to).

5. Don’t allow your analytics team to turn into a reporting one

I’ve seen far too many companies suck the life out of their analytics teams by turning them into report-pullers.  Good analysts bring you insights.  Report-pullers give you data and leave it to you to do the “real” thinking.  If you want the former, you have to guard against the latter.  That’s because it’s really easy to ask for reports from the people who bring you and others analyses.  Tools, clear roles, and an effective manager (more for the VP of Analytics later in this article) can all be components of allowing your people to keep spending the greater proportion of their time doing real value-added work.

Here’s what to ask of your analytics team

For this section, I’m going to speak directly to them, but it’s important that you, the business leader, hear it.  

1. What is the real question being asked?

You are deeper in the data than almost anyone else.  As such, what someone asks for versus what they really wanted versus what’s possible can all be different.  These are some of the better questions I have found to ask before starting the real work:

• Can you (person making the data request) tell me what you’re ultimately hoping to find out or use this information for? Too many times people ask for something specific, thinking they know what it is.  But when you actually get context from them, you can suggest other ways of exploring the issue.  Or you may find something else that no one was initially expecting.  (This question, by the way, when asked from a worn-out team member, can come across as confrontational when it fact it should come from a place of service.)
• Can I repeat back to you what I think I heard to make sure I’m hearing you correctly? This should be standard for anyone receiving a request. Of any sort.  Especially when it can take up a good amount of someone’s time.

2. Make it as easy as possible for someone to “Get it”

I thought this year’s season finale of “Silicon Valley” exemplified this point really well.  But I was so frustrated by what I thought was the real point.  In the show, a tool (it doesn’t matter what it did for our purposes here) was unbelievably powerful, but no one understood how to use it.  It had been optimized for engineers, not consumers.  (In the show, they seemed to completely miss the UI lesson.)

The same is true for analytics.  Hopefully most of the people you’re working with are bright, but they don’t want to struggle to understand what you’ve done. Your effectiveness in your role is in large part dependent on your ability to translate complex analyses into simple, easy-to-understand, actionable insights that other people can understand quickly.  If you can’t easily explain a chart to yourself, get it out of the deck.

Let me repeat that in a different way.  Your job is not to show off every step you took, all the nuances and edge-case exceptions, nor all the technical details of the data.  Your job is to take all those hours of work, ugly spreadsheets and annoying data manipulations and put the end result and suggested next steps onto 1-2 summary pages.

This is what people who are great at their task, in any field, do.  They make the hard look easy.  Ever tried swimming the butterfly stroke or doing a back handspring?  Ridiculously difficult.  And yet, all we see is the end result, never knowing the struggles and challenges others faced.  Same with entrepreneurs, actors or master artists.  And it’s the same with you.

3. Telling a story with the data

It’s not just marketing who is in the story-telling business.  Whether you’re sending out a regular report or sharing learnings from some ad-hoc analysis, what is the value in what you are sharing? Add some color or highlight something that has changed.  Said in a different way – add some real value.  Particularly for ad-hoc analyses, what is the real message that the data has helped to reveal?  Where did you start and where did you end up? Why? What did you come across along the way? Where did it lead you?

4. Look at your subject lines

Examine your own behavior, whether in emails you open or on headlines you click on in social media. You’re much more likely to click when the subject line is hard-hitting, piques your curiosity, or just seems relevant.  If you’ve spent some time on your work or have found something you feel is really cool and actionable, it’s your responsibility to make sure you communicate that effectively.  The first step in doing so, if you’re sharing your results via email, is that subject line.

As a side note, it is unacceptable to say that you didn’t follow up because you didn’t hear back.  Walk down the hall or call the person to schedule a time to review what you sent over.  They may have missed it because your subject line was, well, you know…

5. Know your audience

Everyone processes information differently.  As such, how you communicate to “pure marketers” is likely to be, nay, has to be different from how you communicate to the CFO.  Take a moment to think through the kinds of questions marketers asks versus the CFO.  Those questions reflect a different lens through which they view the same company.  And at the risk of taking the analogy too far, it’s your job to make sure your analysis is in focus for them, not the other way around.

How you present your analysis, as with most things in life, can make a huge difference.  Remember that when you’ve been deep in the data for so long, you know all the nuances, the side analyses you ran, or the way you joined several datasets.  And because of that, it’s easy lose sight of the fact that the people you are presenting to have zero of that context.

So look at what you’re presenting as if you’d never seen it before.  Would someone who has just come from another meeting to look at this understand what’s happening pretty quickly.  Size, design, layout, tables vs. graphs, colors.  Don’t go crazy but these all affect how your analysis is processed.  Learn what your audience prefers and deliver it so that it makes sense to them and in a way they prefer.

By the way, I ESSENTIALLY JUST SAID THE SAME THING 4 DIFFERENT WAYS!! That’s how important presenting information is!

6. Educate the rest of the organization

Take it upon yourself to educate others on the business of the business. Walk them through the LTV model, show them the levers, show them what they can impact, show them why your analysis priorities look like X and not Z.  You have (or you better have) found trends, anomalies, mistakes, or opportunities that can benefit the business.  You need to create the venue to share those learnings.

7. Review your work

On a different note, the best piece of advice I received when I was 22 years old was, “After you’ve spent hours working on something, take 5 minutes to look at it with a fresh set of eyes. And ask yourself, ‘Does this actually make sense? Do these numbers and results seem reasonable?’”  This prevents you from showing gross profit larger than revenues or showing more customers on a subscription on day 60 than started on day 0.

8. Set up internal checks in your spreadsheets

Don’t get too cocky about how good you are, especially the more complex your analysis or model is.  Use some of that skill to guard yourself against mistakes.  For example, you can have a cell that only displays text when your numbers don’t foot.

My favorite story about this was when a partner at an investment bank was going thru a friend’s deck. He got to a page and told my friend, “These numbers are F’ed up.” To which my friend said, “No way, they are right.” Which prompted the partner to turn the page around to show an alert that the analyst had set up but had clearly not looked at.  And in big bold red letters were the words, “These numbers are F’ed up!”  Set up checks.  But again, take a moment to look at what you’re presenting.

9. Stop thinking everyone else has it better

At some point, you’re going to have to get over the fact that your dataset is not perfect.  Some companies have better data.  Some have worse data.  But no one has perfect data. Nothing remotely close to that.  Whatever you think some other company has going for them, I can almost guarantee that there is a good amount they aren’t happy with.

This doesn’t at all mean that you stop pushing for data integrity and better systems.  There are certainly levels for each, below which doing analyses is painfully difficult.  But having seen companies large and small, what I can say is that everyone thinks everyone else has it so much better.  When in fact, everyone is a work in progress.

10. Get over your insecurities

A wise man once told me, “You will succeed to the degree you deal with discomfort.”

I get it.  You aren’t as aggressive as some of the other folks.  People don’t realize how bad the data actually is.  Being a self-promoter doesn’t come naturally to you.  Those senior folks are pretty intimidating.

I certainly don’t mean to cast analysts as weak or feeble.  Far from it.  But I can guarantee that those statements resonated with a whole bunch of analytics folks, just as the first set of scenarios did with the business leaders.

At some point, you have to make a decision to step up.  If you are serious about wanting to help the organization, that’s what it’s going to take.  And certainly, if you are serious about advancing your career, while it’d be nice if everyone were nurtured and developed by their bosses, ultimately you have to take ownership and responsibility for how your career progresses.

A couple notes to the VP of Analytics (assuming you have one)

Most VP’s of Analytics were formerly analysts themselves, where individual contributions were rewarded.  Managing a team is different, but making that transition can be a challenge (again, this holds for many areas well beyond analytics).

And so, speaking to them…

One of the hardest areas to get comfortable with is the idea of training your folks to do something as opposed to doing it yourself.  In the short run, it’s faster to just do it yourself, but part of gaining leverage is the investment you have to make in your team.  That means things will take a bit longer. Will be done a different way (that doesn’t always mean wrong as you’ve hopefully figured out).  And will result in satisfaction and loyalty from your team.

As I see it, you have a few key responsibilities:

1. Make your team look good. As much as possible, give them credit.  Don’t worry, you will get the secondary credit, since it’s your team.  But while you might have been used to hearing the direct feedback about the good work you did, make sure your team is getting that credit now.
2. The opposite also holds – if there’s a mistake, shield them as much as possible. This isn’t to say you don’t hold them accountable.  But if there’s a mistake, it’s in part your job to identify it before analyses is presented.  And it’s not lost on me that this can difficult when you are getting your team up to speed.  But if you throw them under the bus in public, you will lose in the long run.  Trust me on this.
3. Part of your job is to buffer your team from unnecessary requests. This includes requests from the CEO, whose requests can easily get bumped to the top of the list, with almost no regard for whatever else is in queue.  CEO’s rarely filter requests and are known to just send out requests without having a true appreciation of the time and effort involved.

Getting to ROI – A final thought for you, the Business Leader

I know it’s easy to get frustrated with your teams.  As much as that list above is a solid set of areas that you and your analytics team can address, who in your org is ultimately responsible for helping them to get better?

Certainly, it’s important to know where the gaps lie.  If it’s a presentation issue, then look to see who presents information well in your company.  Don’t just look at the “numbers” or finance folks.  In any department, who presents well? Yes, presenting quantitative info is a bit different than presenting creative concepts, for example, but there is still a lot that transcends functional areas.  If it’s one of the other issues, again, think about who in your company does that skill well.

It was intentional that the vast majority of the points geared towards the business side were around philosophical and mindset issues while those for the analytics team were more about communication and day-to-day tactics.  Both sides play a part, as they do in any relationship.  But I’ve also found it’s more effective to approach improving these specific relationships in this way.

That being said, making some adjustments to how you operate is possible and arguably very high ROI.

But I’ll ask the question again – who is actually accountable to develop your analytics team? Meaning, who is helping them how to do their job better?  Who is teaching them how to determine which analyses can lead to something actionable versus ending up in that “interesting but not really relevant” bucket?  And for goodness sake, who is teaching them how to present their analyses in more effective ways?

So that you actually look forward to meeting with them. Because they get you what you want, in a way that you can understand it. And because it leads to something you can take action on.

As opposed to ending those meetings feeling more drained and frustrated than you were an hour earlier.

Just imagine for a moment that at the end of a metrics meeting, you think the following, “Wow, I actually understood what was presented.  It was even more than I had asked for, and best of all, someone else had done the thinking and presented me with a few ideas on how to take action on what we just covered.”

That is not a pipe dream.  That is what a high-functioning analytics organization delivers.  And even if that seems too pie-in-the-sky for you, wouldn’t you rather be closer to that supposed pipe dream than the reality you’re facing today? I’m guessing if you got to this point, the answer is yes.

There is a way to make what you’re doing better.  Maybe there’s an implied judgment in there, but really it’s about acknowledging where you are and then deciding to make it better.

And that’s the sort of decision that you don’t need a metric to help inform.

It’s your move now.

I’d love to hear what are some of the biggest challenges others have faced from your analytics teams?  And if you’re on that team, what is your biggest ask of folks who essentially are your customers?