Expert Spotlight: Gabe Kwakyi on Metrics, KPIs and Predicting Marketing Profitability

In our Star Talent Series, Constellation’s top talent shares their best advice for growing companies. Here in our first installment, Gabe Kwakyi, Co-Founder of Incipia and Felix Analytics and former Paid Search Account Manager at Microsoft, unpacks the difference between metrics and KPIs, and what they can tell you about your marketing profitability.

 

Step into the world of a digital marketer and you’ll probably hear a slew of acronyms being tossed around. CAC, CPC, CRO, CPA, CPI, CTR… whether you’re a walking encyclopedia of marketing terms or all those buzzwords go over your head, a clear and simple understanding of the right way to measure success can ensure your resources are being invested wisely.

 

The good news is, you don’t really even need a detailed understanding of what all those marketing terms mean in order to make smart decisions for your organization with regard to your marketing spend.

 

What you do need is to be able to do is define your goals and distinguish true indicators of success from the numbers that get punched out on your analytics reports… you need to know your Key Performance Indicators (KPIs) from your metrics.

 

Metrics vs KPIs

 

A frequent mistake that many companies make is confounding metrics with KPIs.

Metrics KPIs
  • Cost Per Click (CPC)
  • Click Through Rate (CTR)
  • Cost Per Install (CPI)
  • Customer Acquisition Cost (CAC)
  • Cost Per Day x Retained User
  • Return on Ad Spend (ROAS)
  • Lifetime Value (LTV)

 

Metrics, Gabe explains, are tied to the marketing channel that you’re using (eg. Facebook ads); they should not be thought of as your ‘north star’ performance indicators.

 

By tracking metrics you can compare different campaigns and marketing activities that are operating on a different scale. Metrics can tell you how well one campaign is doing in comparison to another, but this doesn’t necessarily correlate with quality… even strong metrics may not always support your business goals.

 

KPIs on the other hand can tell tell you how your marketing efforts are impacting your goals. These figures are more difficult to parse out, but as Gabe tells, they’re the most useful to track because they allow you to identify the most efficient activities to pursue in order to add value (ie. profit) to your business.

 

The tricky thing is, KPIs don’t reveal themselves instantly; it takes weeks for the data to mature.  

 

That means that by the time you’re able to determine ROAS, you may have spent a lot of money on ads that don’t perform well, when you could have been redirecting your resources elsewhere.

 

That’s where metrics come in handy. They immediately—within a day—give you an indication of the potential impact of a given marketing activity on your business.

 

Say you decide to run a campaign to drive downloads of a mobile app. If you spend $1000 on an ad and 100 people install the app, that gives you a $10 Cost Per Install (CPI). That means your average user needs to spend $10 in order for the campaign to break even.

 

Compare that CPI to other campaigns and you’ll have some indication of the ad’s relative potential magnitude of impact.

 

More pertinently, once you determine how much revenue these users generate on average, you’ll be able to combine these metrics to calculate your KPI—you’ll know how profitable the campaign really was, and this can help you predict which campaigns are likely to be most profitable in future.  

 

A Note on Measuring Data

 

There are many different KPIs that could worth tracking, and several different models you could use to predict performance. A good rule of thumb is to start out by testing multiple models, in order determine which ones are the most useful.

 

Depending on how sophisticated your understanding of data modelling is, you may or may not want to strive for the most accurate model. Complex models are more accurate, but require a lot more tuning and have more room for error—you’ll need some serious expertise.

 

Simple models are somewhat less predictive, but they’re easier to use and will give you a sense of whether you’re headed in the right direction. In some cases, that’s all you need!

 

As your spend grows, the critical question becomes, does your data lend itself well to a simple model, or does it require the kind of complexity that Gabe lives and breathes?