Katie Paine is writing a special series for us titled “How do you measure success?” The Measurement Queen’s series, devoted to examining how PR measurement can be applied to various aspects of public relations to prove their impact, will appear exclusively on the Agility PR Solutions blog.
This is the third piece in the series.
“I can show a direct connection from this Tweet to company revenues,” said no communications professional EVER.
The problem is that in today’s omni-media environment, attribution is complicated, influences are many, and impressions don’t cut it anymore. So tying what you do every day to the bottom line is a complicated process.
It’s one thing if you’re doing social media for an e-commerce company and have sophisticated marketing mix models and CRM systems to spit out your data. But measuring the impact of your social media is a lot more challenging if you’re working in B2B, public affairs or a nonprofit.
But the reality is, the steps you have to take along the way are actually the same.
Step 1: Define “impact”
Whatever type or size of organization you’re working for, there will always be a bottom line—and you will be expected to impact it. So, your first step is to get consensus on exactly what that impact is expected to be and/or what problem social media is expected to solve. If the answer is along the lines of “increase number of followers” get leadership to explain how they expect those followers to make a measurable difference to your mission or business.
Too many measurement programs (never mind overall marketing programs) are based on a set of assumptions that may have been around since MySpace. You need to understand what role (if any) social media plays in the decision-making process. That data is probably with your customer intelligence department if you have one, if not you may need to do a quick survey of your own just to make sure everyone’s assumptions are correct.
Typically, the rationale for social media falls into one of three categories. However, for the purposes of designing your first social media measurement program, we recommend picking one and only one.
- Increase Revenue. This might be translated as “increase market share,” “sell this widget,” or “increase donations or membership.” If you are using social media to essentially push a product or a service, then you are in the “social media as sales tool category.” Leadership that demands that social media impact revenue will only be satisfied if you can show an increase in leads, conversions, donations or some other agreed upon indicator that revenue is or will be coming in the door. That means you’ll need to have access to data from your CRM system (such as Salesforce or Adobe) or some form of conversion data from Google Analytics or whatever web analytics tool you’re using.
- Improve/repair relationships/reputation. Social media is often used to change your image in the marketplace or your community increase awareness of a brand or issue, or reposition an image, i.e., appearing “hip” to a younger generation. If your primary reason for taking the social media plunge is to change perceptions about your brand or your organization, improve your reputation for customer service, or other awareness or preference goals, you’ll need to use pre/post awareness surveys as your impact measure. You could also do a content analysis of the issue you want to be positioned on to see if you have increased your desirable share of conversation around that issue.
Communicate specific information in an educational or emergency capacity. This category is primarily for organizations like the government or the Red Cross that use social media to warn or educate people. Typical impact metrics for this use of social media would be the percentage of the population that changed their behavior or behaved in the desired manner—i.e. got vaccinated, stopped smoking, evacuated, went to a shelter, or contacted the correct emergency responders.
Step 2: Ignore your dashboard
If you’re using a commercial social or digital media dashboard report that comes out of the likes of Hootsuite, Sprinklr or Simply Measured do NOT simply use the metrics they push out. Start by downloading the raw data into an Excel spread sheet. Ignore any invented algorithms or metrics like “engagement” unless their definition is in perfect alignment with your own internal definitions.
Ignore any metrics that sound like “potential impressions” or “potential reach”—they are pure unadulterated fiction. Likes belong in the same trash heap. Given the prevalence of bots and fake followers, these numbers are about as likely to be accurate as your average politician in a campaign speech.
The problem is that the good people who design such tools are not intimately familiar with your organization and how it defines success so you need to analyze your data for the specific metrics you’ve agreed upon with senior leadership.
Step 3: Validate your data
There are dozens of ways that your data can get corrupted. Your search strings may have included irrelevant data—i.e., information about underground trains (Subways) if you’re a prominent sandwich chain (Subway.)
One of the most common data problems is if you are using multiple agencies and platforms and don’t have a common dictionary of terms. Make sure everyone means the same thing. For example, if you are using “click thru” as a metric and are using the data from Twitter Analytics or Facebook Insights or other similar data provided by the platform you’re using, check it against your own internal data from Google Analytics, Web Trends, Omniture or whatever you are using. If it doesn’t match, toss it out and use your internal metrics. There is no faster way to lose credibility with management than to have data that is inconsistent with the other data they are looking at. Internal web analytics are considered more credible than any platform data given the self-interest of those platforms.
Step 4: Define a benchmark
If you’re just tracking our own social and digital results (no competitors) you will need to establish a baseline against which to judge future performance. We recommend an average of the prior 6 months (or a minimum of three if you don’t have data for six).
Better still is track your performance against your competition or the opposition (it’s whatever keeps your CEO up at night.). The best measurement platforms can easily track competition brands and that way you have a clear view of what else is going on in your industry.
Step 5: Define your Key Performance Indicators
Key Performance Indicators (KPIs) are the specific metrics that will define your success. They typically start with a percentage, as in:
- Percentage of discussion in which we were favorably positioned on the issue of diversity in the workforce.
- Percentage increase in website visits lasting five or more minutes.
- Percentage increase in registrations.
- Percentage share of favorable positioning in the marketplace.
- Percentage share of desirable discussion.
- Percentage share of quotes.
- Percentage increase in awareness.
- Percentage increase in willing to consider or prefer.
Whatever metrics you agree on, be careful and choose wisely—you become what you measure. In other words, you should measure whatever you need to do more of. Many organizations chose to combine metrics into their own “social media engagement index” or “social media quality index.” You can find instructions on how to do that here:
Step 6: Analyze the results, draw conclusions, glean insight, and present your report
All your data is just trivia unless you analyze it for insight and meaning. Who cares if you’re killing it on SnapChat? How does that affect your qualified lead volume or your next budget decision?
Gather your data, separating the desirable conversations from the undesirable. Both are meaningful but you can’t lump them together because (obviously) the social impact will be different. One key metric is to ascertain the degree to which you’re getting your fair share of the good conversations and avoiding the bad ones.
Don’t forget to look at search rankings and results. If you’re using Google analytics, you can easily track the percentage of searches that are generated by social media and track their contribution to conversions over time.
Look at repeat visitors, not just visits. Chances of getting revenue or recommendations out of repeat visitors is much higher than one timers. Besides, a single visit may well be generated by a bot or a random false search, so focus on those actions that are most likely to be generated by your target audiences.
Remember: It’s not about the shares, it’s about the endorsements. When you are measuring “virality” or sharing, what matters is not just the share itself, but whether there is any commentary or augmentation of your brand or message. And whatever you do, don’t count your own shares, that’s cheating.
Step 7: Learn from the experience
Finally, remember what senior leadership really wants to know—what didn’t work and how can we improve. Now go through your data again and decide how it will answer those questions. That’s the story you need the data to tell.