Analytics Mistakes Marketing Agencies Make

There are many mistakes that marketing agencies make when reporting on their performance and optimizing their campaigns. These mistakes range from inefficiencies to wrong assumptions being made. They can be tiny over-sights or large and can ultimately cost you your clients. As a marketing analytics agency, we are often hired to help identify issues with current reporting, marketing team’s inefficiencies and opportunities for improvements. What we see over and over again is that most of the mistakes are not due to malicious intent but rather due to lack of education or awareness on how reporting works. Here are the most common mistakes that marketing agencies make and how to fix them going forward.

Descriptive metrics versus insights:

How often do you report to your client we had 10 million impressions from TikTok Ads this month, we had a CTR of 2% or we had 10 leads this month. All those examples have one thing in common they are descriptive in nature, there are no insights available. It is equivalent to saying it is raining. The more insightful or predictive analysis would be it is going to rain for the next few hours, or saying rain is finally helping forest fires stop. Therefore, for marketing agencies a more insightful metric would be to say something in the lines of we had 10% more impressions this month compares to last, our CTR is better then our benchmark and we had 10 leads, 5 of which qualified, which is double the amount of leads we received last year.

So how can you fix it: ask your analyst or your marketing analytics agency what does this metric or graph tell me and what questions does it answer, is it the full picture? This will help narrow down what you are reporting or change how you are reporting these metrics. The second thing to do is ask what is missing? Does it give me a full picture, or ideally we would also figure our the why? Why are impressions going up, CTR is above benchmark and what campaigns contributed to your leads are all great starting point, you may want to show impressions month over month by campaign or influencer, your CTR by ad creative and your leads by channels to really understand the impact of your marketing so you can have predictable results in the future.

Not doing Attribution Models

What are attribution models? They are models that allocate where you lead, sales or another type of conversion came from. The models can be based on first touch, last touch, linear or time serious. Online tools such as Google Analytics (GA4), Segment or similar tools are able to sell you on their attribution capabilities some even in real time. While GA4 is free, others are not, and what happens if your marketing agency is running offline channels such as TV or Radio, or your marketing manager forgot to put UTM codes or GDRP is making it impossible to track a big percentage of your website visitors?

So the bottom line is your attribution is only as good as your data. Limitations in data collections can make it tricky to build statistically significant marketing attribution model. However, all is not lost. You can run decision tree models or regression models to identify sources that not only contributed to your conversion but more importantly what sources in combination played a role. What is even better, is that your data maybe limited but you can still see trends. For example, a dealership may have 10 people walk into their showroom on a Saturday morning. You may have your sales reps ask how did you hear about us? And you may get answers such as my friend bought a car here, I live next door, I saw a commercial or simply no idea. Not very useful, but what is useful is being able to say: 10 people came, we ran tv ad, radio ad, social media campaigns and search ads, each one of these sources has their own success metrics such as impressions, clicks, GRPs and so on… by apply it to the model and using more than just Saturday morning’s data, we can see trends of when something was running and was not and its correlation on number of people coming into the store. While it may seem complicated your marketing analytics company can help identify sources of conversions and if they cannot advise you on what testing you can do in order to achieve the attribution model you are looking for.

Not Testing Everything

How many agencies have launched nurture campaigns without testing how many days between emails they should wait prior to launching a new email? How many agencies do creative tests prior to launching their main campaign? And how many marketing agencies use a testing matrix when doing their budgets and campaign planning for the year? The answer is very few! This is a huge opportunity, each test can yield insights to improve results. This way when you have your monthly meetings, you can say things like ‘We tested an ad with a Portuguese Water Dog and a Pug, and the Portuguese Water Dog generated higher CTR and clicked more on ‘where to find’ button.

What is even more helpful is your team can stop arguing about subject lines, concepts and messages. Why? Simple, by putting all ideas to the test, after all there are no bad ideas! A way to run tests is to either use tool’s A/B testing or run your own small test for a specific time frame and record your results in your testing matrix. As long as your results are documented and are on a smaller scale but yet statistically significant then you are almost guaranteed to see improvements over time. Ask your marketing agency to help you with setting up your testing matrix, helping identify if you just need simple A/B testing or a more sophisticated multi-variant analysis, and reading your results with statistical significance.  

Segmentation Model

Segmentation simply means grouping different people into like-minded groups. For example, if you are selling tennis and golf shoes you may want to segment your customers into 3 groups: tennis group, golf group and tennis & golf group. This will ensure that you are sending the right message and creative to the right group. What we see with analysis is that when you do this your performance usually increases. Why? Because your marketing is relevant, and if its irrelevant then you will lose your audience over time. For B2B Marketing agencies segmentation is often done by title, industry, events that the leads attended, piece of content they consumed and so on. For B2C marketers, they can segment based on demographical information (although due to privacy laws this is more difficult to do), and more importantly based on user’s/customers behaviour. For example, if a person has clicked on an ad for shampoo for curly hair they probably have curly hair and would require other curly hair products. You can also segment based on consumption behavior. For example, you may realize that the majority of your customers are on Pinterest and they pin images of plants therefore, you may decide to run more ads there to find look-alike groups to sell your plant pots.

Not Automating Reports

The worst inefficiency a marketing agency can do is not automating reports. If you have account managers still trying to do budget pacing in excel, your social media team copying and pasting creatives into PowerPoint every month, or your marketing analyst pulling data from multiple platforms just to report on total marketing spent and MROI, then you need automation. Let’s put it in other terms, would you trust a stockbroker who still phones in his trades? Or what about a coffee shop that still grinds their beans by hand, and brews espresso on a cooktop. Ok so maybe the last example, can have some nostalgia and romance to it, if you have 30 minutes to wait, chances are you prefer Starbucks prior to your 9am meeting. Therefore, it is advisable to stop doing manual reporting, when automation will save you time and money!

You can hire a marketing visualization expert to automate reporting, they can identify ways to blend/merge data sources and report on results in a meaningful way. There are many tools on the market that will do just that from free or almost free Google’s DataStudio (Looker Reports) to a more expensive tool such as Salesforce’s Marketing Intelligence cloud (Datorama). Your marketing analytics partner can help choose the right tool for you!

Outsourcing Marketing Analytics

One of the reasons why marketing agencies have not automated reports, do not do continuous testing, have segmentation models and attribution models is due to lack of expertise inside or simply lack of time. Other agencies will comment on things like privacy, which can easily be overcome with NDA and hiring a reputable company in North America. Another argument is high cost, typically a cost upfront can save hours, improve efficiency and more importantly stop you from losing clients. There is very little downside to bringing in outside marketing analytics help.

About StrategicDB

StrategicDB is a full-service marketing analytics firm. Specializing in working with marketing agencies for the purpose of making them look insightful, data-driven, and efficient modern marketers. We pioneer marketing testing, utilizing latest technology with the time proven methodology to help marketers become more efficient and driving better results for their clients. Contact us today to learn more about what we can do for your agency.