Trust.
Trust is a tricky character. We all want people to give us theirs, and yet we are super wary of giving it out ourselves. In family, in love, and of course, in business.
Earning the trust of others as a data analyst is a tall order. Trust is an ephemeral entity that can instantly vaporize during a presentation if certain checks and balances aren’t in place to cement your credibility.
At the end of my flagship Raise the Bar Chart Boot Camp workshop for analysts and marketers, I conduct the Live Exercise, where students present a revamped presentation to the class using the principles I teach. The students then communally critique the presentation with my gentle assistance. The Live Exercise is arguably the most valuable component of the training because it locks in core concepts in a way that teaching alone can’t accomplish.
And after delivering this workshop enough times, I’ve identified patterns of how we unwittingly present data in a way that can undermine our audience’s confidence in us.
This post will take you through the five most common mistakes we make in presenting data that break trust.
Now, we’re going to have a little fun with this one. I’m going to show you each slide first, and then ask you to take a minute to be your audience and think about what might throw you off. Think of it as a modified “Check…and Double Check” exercise of Highlights Magazine yesteryear.
Alright, you game?
Of course you are! Here we go!
Challenge #1
From an audience perspective, what would throw you off about this slide?
ANSWER: The title insight doesn’t match the visualization
Notice that while the title insight references campaign-level performance, the graph only depicts what’s happening on the channel level.
A mismatched insight to visual can create a trust chasm between you and your audience. Most accusations won’t hold up in a court of law without presenting the evidence. Presenting your hard-won insights to an audience of skeptical stakeholders is no different.
Visual reinforcement of your statements not only can bridge the trust gap, but research suggests that it can also increase the recall of your information up to 65% three days later [source]. So if you’re using a chart to substantiate your claim, make sure the written insight is easily gleaned from the visual. Here’s how I’d present it:
Notice how I’m reinforcing what the graph is showing, but also providing calculated context within the portfolio.
One of the most common reasons I see this happen is because of what I call “crystal ball” syndrome: the analyst has access to all the data things, and quotes a stat they came across during an offline analysis. But that stat isn’t reinforced by the visual, so it seems to the audience that either the presenter is eerily clairvoyant or just full of…well you know.
The exception I use to this rule is when I’m using just text and relevant imagery to visually communicate a stat or insight. Here’s an example:
Note that I use this technique sparingly, and often I follow up that slide charting a related metric that visually represents both of the insights. Or, I’ll include a charted representation of the figure in the appendix.
BOTTOM LINE: The critical role of your graphical visuals are to reinforce your message, so take care to make sure your insight and your visual are on the same page!
Isn’t this fun?? Ok, on to the next example:
Challenge #2
ANSWER: Using subjective words to describe performance
I see this one a lot. A lot a lot a lot. “Did better” is maaaaaybe not the most impactful way of phrasing what exactly went down here because it sounds subjective.
Subjectivity is a tricky character in a setting where neutrality is key in facilitating data-driven decisions. Subjective assessments can look like:
- “Q3 clickthrough rate was worse than as Q2”
- “Creative A didn’t perform as well as than Creative B”
- “Mobile search conversion rate was not as good as desktop conversion rate”
And yaddah. Note that these are all statements I’ve seen presented during performance readouts. It’s important to be mindful that words like “good”, “bad”, “better”, and “worse”, are subjective judgments.
Nicely neutral descriptors for performance and comparison include:
- Higher / Lower
- Increased / Decreased
- Outperformed / Underperformed
- Elevated / Declined
- Exceeded, Outpaced, etc.
Now, that’s not to say that you will always have 100% statistical confidence in your statements, or you may get asked a question on the spot where your best guess will have to suffice. Here’s how you can represent your story with greater dispassion:
BOTTOM LINE: The more concrete your language, the more trust your audience has in your insights. Avoid subjective assessments when possible.
Oh and, that Empire hover cats beat basic hover cats. Every. Single. Time.
And that segues ever so nicely into our next pitfall. Well, what have we here?
Challenge #3
Don't peek at the answer!! Ok, I’ll give you a hint – it’s related to our previous oopsie daisy:
ANSWER: Vague performance language without context
Just saying a metric increased isn’t enough; your audience will instantly want to know by how much.
This is another common pitfall that won’t necessarily break trust, but it falls short in communicating as much pertinent information as possible to your stakeholders. When presenting data, you want to eliminate the opportunity for as many question marks as possible.
Observations that create more questions than answers include:
- “Conversion rate went down last month”
- “Keyword A performance was higher than Keyword B in Q3”
- “Online sales exceeded our benchmark”
From my decade of experience observing and speaking to stakeholders about data, I notice a strong propensity for them to demand more contextual assessments.
Always remember that one of your most important goals as a presenter is for your audience to do less work by answering their questions before they ask them.
Here’s how I would rephrase this observation to slake my stakeholder’s thirst for context:
BOTTOM LINE: Context is key in creating trust in your work because it shows you’ve gone the extra mile to anticipate a likely question they’ll ask about performance.
Alright, next example!
Challenge #4
This issue falls once again into the “prove it!” camp. And the pitfall here is…
ANSWER: Referencing targets or goals not visually displayed
Targets and goals make the business world go ‘round. Without them, we have no sense of accountability for continuous improvement. I can feel question marks arise from audience members’ heads as soon as a goal is referenced without being clearly articulated.
Once again, those questions marks translate to extra work your stakeholder’s brain is doing to create context and keep up with you. I prefer to add a target line in black for total neutrality, but use a dot or dashed style to indicate that it is a projection, not actual data. As so:
BOTTOM LINE: Context is key in creating trust in your work because it shows you’ve gone the extra mile to anticipate a likely question they’ll ask about performance.
And last but by no means least, our final challenge:
Challenge #5
Tricky one, eh? What could possibly be missing? I’ll give you a hint: it’s been present on every example except for this one. One itty-bitty teeny-weeny detail (where the devil liveth):
ANSWER: Not citing your data source
This is an easy fix for preventing loss of trust. Every single time a numerical value is expressed in a presentation, a data source MUST be cited. This is one non-graphical element that doesn’t count as extraneous “slide fluff” like logos and watermarks and is essential for engendering analysis trust. The format is as simple as follows:
“Source: Platform/Reference, Month Year.” In our example it makes a little cameo in the bottom left corner like so:
And, the oft-cited rule for a minimum font size of 20 pt need not apply since it doesn’t articulate a story-specific message. I usually set the source disclosure to around 10 pt, depending on the font.
BOTTOM LINE: Context is key in creating trust in your work because it shows you’ve gone the extra mile to anticipate a likely question they’ll ask about performance.
Final Thoughts
Building trust with your stakeholders during data presentations is an essential component to using that time productively to inspire action and create your indispensability. I’m confident that with just being mindful of the most common trust-breaking pitfalls, you’ll be winning your clients and execs over every time you present.
PS – For the love of all that is holy moly, please do not start a presentation with, “So the bad news is…” This is an excellent way to trigger your stakeholders and put them on high alert for negativity. I believe that there is no bad news in analytics insights, only opportunities.
Campaign missed a target? Opportunity to present new ideas for meeting it next month.
Publisher underdelivered on a campaign? Money back for a make-good that can be reallocated to a more effective channel.
Data feed failed? Resolve it immediately, estimate a timeframe needed to collect enough additional data for analysis and accept that we are not gods of technology.
For me, part of the art of presenting data is about reframing “bad news” into opportunities, not failures.
You only fail if you fail to take action on what you’ve learned.
Put that on a t-shirt!
