Start to measure performance of campaigns through conversions which extend well beyond first-click or last-click models. Return on investment takes time, don't expect your analyst to be able to provide a real-time dashboard in which you will see the ROI of each tweet.
At most company presentations I see:
- Data that only confirm prior beliefs
- Strictly positive performance of the division being represented
- Blanket statements about customer demographics
- Marketing campaign plans without goals or objectives
- Statistics that aren't relevant to the audience
**Warning, beginning of a rant**
Do any, some or all of those sound familiar?
These tendencies are (usually) not intended to mislead people. Simply put, a lot of people who actually present information tend to dictate what they would like on slides rather than asking their analysts to provide insights and data.
So, what is going on here?
The problem isn't typically the top down approach to presentations. There's a certain degree of direction that a presenter must have in order to get a point across. A presentation without clear direction is similar to a website that is designed by committee - it has a higher tendency to be ineffective.
However, there is a disconnect between presenters and those who are supplying information. Many people who are being asked to provide data aren't clued into the context of the meeting. A web analyst who is being asked to "provide year-over-year traffic growth categories" becomes immediately handicapped - being forced to find only positive results. What if the growth categories didn't seem very relevant but the stagnant or decaying categories had great potential? This information, while not intended, is not actually an indicator of business performance.
If an ad campaign finishes and a presenter wants to show off the results, he or she may ask for performance data but then cherry pick topics which look great. Some people even ask for "expectations" of the campaign after it has already concluded - creating another (unintentional) form of misleading information due to lack of forecasting, goal setting and planning.
The list goes on...
How can this be fixed?
First of all, presenters need to sit down with analysts and explain who the audience will be along with what decisions will be made after the information is presented. This is absolutely critical to driving successful analysts. Secondly, bring analysts into more "high-level" meetings. A good analyst will be able to listen to the problems the company is facing and start to provide data analysis and insights which people may not even know exists. Keeping analysts in the dark and then expecting a data dump will severely hinder the ability to get actionable insights. Thirdly, encourage analysts to speak up and ask questions. In many organizations, a "need-to-know basis" exists in which people may not feel comfortable asking questions which are "out of their pay grade." Do not let that happen to you!
**End of rant**
I read a post by Bob Hoffman, the writer of Ad Contrarian, which had a simple but powerful sentence that resonated with me.
I hate to break it to those who are in love with the industry, but if we look at a brief history of advertising, it becomes apparent very quickly that the overwhelming majority of people simply do not like it.
Here are a few examples of advertising vs. response:
Television -> DVR, Tivo, etc.
Radio -> Podcasts
Display banners -> Ad blockers
Native Advertising -> ...Frustration...
Social Media (Paid) Advertising -> ...Frustration...
I use the word frustration because these are areas in which people are working on solutions similar to those in TV, radio and display. Here is a cherry-picked chart to illustrate the exact point I'm trying to make.
In the near future, you will have to decide if you want to be an organization that:
Is increasingly dependent on the ability to track individuals through new and innovative ways.
Is looking for the next cutting edge form of advertising which will disrupt the ad industry!!!
Is less dependent on ad tech and is actively creating a fan base by offering products and services which people can't stop talking about.
Fitting into the #1 category means you will be using some of the latest and "greatest" ad tech. You will be constantly trying to interrupt a massive population (which is still growing) from what they're doing online.
If you are in the #2 category, you're probably delusional. Someone in this category once told me, "the pop-up ad was the greatest invention of its time" because while "it is annoying, it got your attention." I don't think I need to comment on this.
The #3 category definitely requires the most work but will have the largest payoff. There is no clearcut road map for success - otherwise everyone would be doing it! As all of us in the marketing industry know, longterm success isn't based off of how cool your advertisements are.
I'm happy to speak about measurement strategies with the aspiring #3 people! The analytics are simple to understand but tricky to implement into an organization.
I hate to be the one to break it to you, but the average customer shouldn't be that important to you. I'm not writing this to repeat the marketing rhetoric you hear about how "millennials" want massive amounts of choice, personalization, customization, etc. I'm here to tell you that people misunderstand the facts about their business - even when the numbers are correct!
Recently, I was asked to look at marketing performance for a company in Chicago. The company told me, "on average, we make about $100 per transaction." At first glance, they appeared to be spot on.
Daily averages showed a nice normal distribution with a mean of $100.
With that knowledge in mind, the company had set a hard and fast CPA goal which could take into account profit margin and all of the other goodies! All of the marketing tactics and materials were developed and targeted for the average customer at the average price of $100.
Sounds great, what's the problem?
First of all, transactions are different than customers. One customer can make multiple transactions! Let's put that point aside for now and assume each person is only buying once. The company's numbers weren't wrong but they didn't accurately describe what was going on. Take a look at the chart below:
The company offers two different ways to purchase their products, online and in-store. Clearly, there is a discrepancy between the shopping behaviors. Those who are buying in a store are spending an average of $110 and online the average is $90 per purchase.
Now you have two averages rather than one, is that actually better?
With the technology we have today, it is easier now than ever to figure out what creates these discrepancies and how to make the most out of them.
In the case of this company:
- Every online purchase was given a 10% discount if the customer signed up for an email newsletter
- Discounts through affiliate programs were attached to the majority of online orders
- In store purchasing capitalized on upselling techniques by salespeople
- Stores had a prominent display of multiple pricing options, which created larger variance compared to online (which should help merchandisers)
Now the business can approach online vs. offline marketing with a bit more knowledge, so I would say it's a great thing to have two averages!
Keep in mind the average customer is not a real person. While you can tailor your marketing to this group, make sure you're not wasting your money. Drill into your data where necessary but don't go too far. Knowing where to slice and dice takes skill and effort - don't try this at home!
I'll save the changes in strategy for another post, thanks for reading!
**FYI** In this post, the story is real but the data is simulated (for the sake of privacy). While real data is always much messier, I hope you'll find some of this useful.
I will be giving a presentation on data visualization this Thursday (April 13th) from 4:30-5:30pm at 242 Linden St. Fort Collins, CO.
Please come down and enjoy yourself!
You can expect a fun learning experience with a chance to meet some great people! I will gloss over the history of data visualization and give you some tips to improve your work.
There are a lot of examples of what not to do out there (see chart below) - don't get caught doing something like this! :)
The details can be found here: http://conta.cc/2oWVLpU