How Creative Agencies Can Use Analytics to Prove Success
We sat down with Mailchimp to talk about how we use analytics to prove success.
1. Tell me a little bit about Jellypepper - the type of work you do and the type of clients you work with
In short, we're a lean creative agency. We work closely with startups and tech companies to turn bright ideas into sleek products, brands and websites. Our work so far has primarily been with what we're calling "disruptive startups" — tech startups in fields like self-driving cars, artificial intelligence, drones, 3D biotechnology and cybersecurity.
2. Do you consider analytics and measurement an essential part of how your agency markets itself to new (and potential) clients?
Absolutely. Our proposal pitch deck template focuses on a few key things (past clients, testimonials, the team) but one of the key selling points is the case studies and the accompanying analytics. We surface changes in key success metrics as indicators of how effective our methods are.
A lot of the time, when agencies pitch and promote themselves, it's based on promises and relationships. We try to focus more on proof and analytics are one of those key concepts that help us present this. We're currently sprucing up the case studies on our site at the moment so you'll see some success metrics there soon!
3. Do clients today expect you to be tracking metrics on the work you create for them? What are the benefits of this for clients and agencies?
End-to-end, cross-platform analytics across the website and products we build are crucial for startups to measure the effectiveness of their new user journeys and prove their potential to investors.
For us, not only is it a proof point that our work is effective, but it also helps set the benchmark to improve! For example, if we see the conversion rate for the new website has moved from 10% → 20%, that means there's a whole heap of experimentation and work we can do with them in the long-term to bring it up even further.
4. What type of metrics do you typically track for clients? How do you decide which type of metrics make the most sense for each clients?
It really depends on the type of company and what their goals are, but from a baseline, we aim to track pirate metrics (Awareness, Acquisition, Activation, Revenue, Retention, Referral) through high-level behavioral events across the website and product, as well as high-level behavioural metrics for emails. We focus on the first four, as retention and referral can be a bit harder to track, particularly as an agency as you might be on the project for a limited time.
5. Do you think agencies spend enough time focused on gaining insight on how client work is performing by analyzing metrics?
I think it can be difficult as a lot of project-based agency work has a hard start date and deadline and sometimes the relationships aren't as tight as you'd like them to be after the project, making it hard to follow up.
It can also be quite daunting for radical projects, where you're rebranding and redesigning an entire site. The short-term analytics might take a dive because people aren't used to the new style (despite an improved UX) but those metrics might improve past the original in the long-term. But you'll only get to see the short-term as an agency.
That being said, I think as companies become more growth-focused and rely on data as a way of proving whether something is good or not, agencies will need to respond accordingly and start taking an interest in the long-term success of their clients and projects through analytics. Positive analytics are great for the agency and the client — everyone wins.
6. Is tracking analytics on client work easier than some agencies might assume?
Absolutely. For web and products, it's essentially 3 steps:
- Identify which metrics you want to track and how you measure them. You can base this on a user journey, likely points of churn or, of course, your standard pirate metrics (AAARRR).
- Identify what platforms are best for collecting and visualising these analytics. We recommend Amplitude and Mixpanel, as well as Segment to manage the pipeline. If you are working with an early-stage company, many analytics tools have very generous free or startup-focused plans. Have your clients set up the platforms of their choice and send you the API keys or account details.
- Add an analytics call at each of those key areas in your codebase.
The third step can be quite large when you're dealing with a complex app or website. Email campaigns and newsletters (particularly with Mailchimp) are a lot easier to set up as they come with detailed behavioral metrics, demographics, engagement and segments automatically. ❤️
7. At what point in a project should you begin tracking? What is a good place to get started?
The sooner the better, especially since good tracking requires a bit of strategy before implementing. It's likely that the metrics will be a bit distorted when you first launch your product (the first week or so) but after that, they'll even out and you'll start to see more regular patterns.
You don't need to dive into tooling and code to get started! Just open up Notion or Word and start answering questions like:
- How you define success in your product?
- What does "activation" mean in your funnel?
- Do we have a solid referral system in place?
Or if you're using Mailchimp, just send an email 😉
8. Do you typically analyze metrics for different audiences for each project? Different channels? How do you decide what metrics are best to track for each channel/project?
We typically analyze metrics for different audiences with different platforms based on the goal. If you're looking for high-level audience information like a users' country of origin, type of device and how they're moving through your site, Google Analytics is a brilliant go-to tool. If you're after something more custom (how different cohorts are interacting with your content based on traits and low-level demographics), we use tools like Amplitude.
For each channel or medium, it's important to track a mix of qualitative and quantitative metrics, as well as identify which metrics are going to answer the questions you're asking. Different channels have different methods of answering questions — engagement on email is measured by open, clickthrough and reply rates where as it's all about likes and shares on social media. There are a few metrics you should track for each channel with standards based on your industry, but beyond that it can be tailored to your product or service!
9. Do you work with clients to determine their goals for each project, so you know what to track?
We run workshops at the start of a website or product project to define the key metrics and what success looks like at the start of the project so we're able to measure it effectively. We centre the conversation around their business goals (conversion, awareness, public perception, etc.) and what tools they'll need to achieve this.
10. What are the most important metrics to track on different channels (email, social, website, direct, etc.) any why?
When it comes to email marketing, important metrics include:
- Behavioural metrics (open, clickthrough, unsubscribe and bounce rates). You can find this on the report summary for a particular campaign. This helps indicate the quality of your content. Tracking this through to website or product conversions will show the ROI of your email marketing.
- Engagement cohorts (which groups of people are opening your email often, sometimes or rarely?). You can find this on the Mailchimp dashboard. This will show you who which audiences you should be focusing on and who's finding value in your content.
- Audience characteristics (industries, sizes and demographics). You can find this on the report summary for a particular campaign. This helps you understand who you're talking to and what they have in common so you can tailor your campaigns.
For social media marketing, you can focus on:
- Engagement (likes, shares, faves). Shows how responsive and engaged your audience is.
- Awareness (impressions and reach). Shows how many people you're reaching, good for tracking brand awareness and influencing public perception.
- Sentiment (reactions, comments). Shows how people are responding emotionally to your content or business.
- Conversion (all the 3-letter acronyms): when you know the above, you can start to measure the financial aspects of your efforts: return on investment (ROI), cost per click (CPC), cost per acquisition (CPA), click-through rate (CTR) and cost per thousand impressions (CPM).
Website design is a bit more custom, but it's handy to start by focusing on pirate metrics (AAARRR) and your North Star metric (the metric across your business most predictive of long-term success). From there, it depends on your user journey and business goals.
Direct mail is an interesting one as there's no real-time analytics you can collect. However, methods like Response rate or Conversions (through Promo / QR Codes or URLs) can pave the way to measuring the more common financial metrics such as cost per acquisition or customer lifetime value.
With OTT or CTV, you're essentially tracking video or ad analytics. For the video side of things, important metrics would include video completion rate (or identifying common churn points) and with ads, attribution tracking. With the rise of low-cost subscription OTT streaming services like Amazon Prime and Netflix, this will be an interesting space to watch.
11. Any tips for presenting data in a way that is easy and digestible for clients to understand? How can you use data to help clients see what's working and what changes are needed, as well as highlight what a great job your agency is doing?
I'd recommend starting off by choosing a platform to help you collect and interpret this data. Amplitude, Mixpanel and Heap are great for websites. For email, check out Mailchimp's reporting feature.
From there, I'd figure out what you want to convey to your client and create a presentation that features some of these graphs and key metrics, drawing attention to specific points of interest. If you're focusing in on a particular metric, you'll want to be careful that your graphs aren't misleading.
A really cool option for the pros is to stash all your cross-platform analytics (from your emails, website, products, etc.) in a data warehouse like S3 / Redshift and use a tool like Metabase to start asking your own questions.
1 July 2021