Data headaches run in all organisations

At a Marketing roundtable lunch last week, I learned three things.

1. The burrata salad had great consistency. The outside is firm but not dry. And on the inside, runny but not watery.
2. Rocket and tomatoes help cut through the richness of the cheese. 
3. Data headaches run in all organisations.

I have a love-hate relationship with data. I enjoy slicing things up and moving things around to see what we can do to improve our conversions and retention.

But sometimes, it feels like trudging through muddy waters. Sometimes, there’s not enough data to deduce anything. Other times, it’s hard to make sense of it all.

As the restaurant served lunch course after course, we surfaced three main problems with data.

First, it’s hard to trust the data

With humans keying it in, both on the customer and employee side, there are bound to be errors. 

Names misspelt. Phone numbers entered wrongly. And even if there are dropdown options, people could still select the wrong choices.

More. If email addresses are wrong and emails bounce, are they reconciled? If customers request to update their details, are they updated on the system? Are duplicate profiles merged?

These are just a fraction of the problems.

We can only do our best to maintain the data’s cleanliness and accuracy.

For example:

1. Converting short text fields to dropdowns as much as possible
2. Using data validation tools on forms
3. Conducting regular sales training and reminders to follow best practices

P.S. During the session, I also learnt that some people have so many trust issues with data that they tear down reports and rebuild them themselves. Isn’t that a lot of time wasted? 🧐

Second, there’s not enough data

One example brought up was pre- and post-Covid. Behaviours changed drastically. The pre-Covid data was no longer valid. We’re starting from scratch.

Next, what if the businesses we’re targeting are big and few? How can we say that this persona behaves this way and, therefore, we should tweak our message a certain way?

Last, of many, we have the attribution problem.

The easiest way to attribute a lead is the last-touch model. But no customer would see an ad and buy right away. A hypothetical journey could look like this:

But we don’t have those data.

So if Facebook has the highest CAC, do we scale it down? If an influencer doesn’t convert, do we stop working with them? 

It makes sense to move to a multi-touch attribution model, where we attribute a sale to multiple channels.

That is complex and involves a lot of resources. It requires the organisation to change its mindset and stand behind it to move things along. But I think it’d be an exciting project to work on 🙂

Third, there’s too much data 

Every number plays a part in bringing us the results. But too much data presented, results in analysis paralysis.

To move fast, we need a setup that tells the story in 5 minutes. Then, when we see a potential problem or opportunity, we dig deeper.

Here’s a good way to structure data.

Level 1 Data: Data that is aligned with the company KPIs. These numbers tell you the “What?” What’s happening to the company?

Level 2 Data: Data that directly impact level 1 data. Level 2 Data tells you the “Why?” Why is the CAC so high?

Level 3 Data: Numbers that indirectly impact the organisation, like Instagram followers, email open rates, etc.

There’s so much more to learn in the world of data, and I’m just scratching the surface in this brief reflection. I’ll be back to share more learnings as I go along 🙂

Also, lunch was great. But I’m curious: Why do European meals always end with coffee or tea? Would it be to counteract the sleepy effects of a heavy meal? Yawn…