What Is An Attribution Model For Startups?
Lets talk about attribution today. This is not going to be a post on how to solve attribution as there is more then one way to do it. This post is about figuring out how you can think about attribution for your brand. Being a startup on limited cash and people can make attribution a challenge. Below is what I found to work as an alternative option.
Video Transcription Below. Enjoy and don’t forget to leave your comments below.
My name is Duane and hello from beautiful Vancouver. You may remember me from such conference stages as MozCon, BrightonSEO, HeroConf and SMX.
Today I want to talk about Attribution Modeling – the how we think about it and why we do it for startups that just launched. As we anything we do at the company. Test and learn is important.
The startup marketing framework I have been using below has been refined over the last 6 years. It’s not perfect, though it’s a starting point and good is certainly better than perfect so we can launch something and iterate as we go. If you’re a startup and resources are scarce, than you are in a boat I have and continue to be in as we work we more startups at the company.
Today I am sharing the framework I’ve worked on and what shape it has taken. To build this framework I started by answering the 3 questions:
- How Will We Define Success?
- What Is Our Attribution Window?
- What Should Our Framework Look Like?
Having a customer sign up for your service or buy a product is usually the main key performance indicators (KPI) . However, I also look at people subscribing to the blog, downloading an ebook or whitepaper. The latter two are tracked because a campaign you thought would drive customer acquisition, is actually more suited for blog subscribers or getting people to download an ebook. If your startup offers webinars or product demos and you’ve those as goals (or events) in Google Analytics, than you can add that to the spreadsheet below.
I use to pick a 30 day attribution window for the cookie that sits on a customer’s computer. However, when I think about the effect of a campaign and the value it provides to a startup. I started to pick 60 days because most startups do not have long sales cycles and if someone takes longer than 60 days to convert, it becomes a mess to track. Plus Google Analytics (GA) defaults to 60 days and having all your advertising and GA sync up makes the job easier.
The spreadsheet below is what I created when I start to assemble all my data for different campaigns we ran. I added in a threshold of $500 / campaigns because anything under that limit doesn’t tend to have enough data to make a proper informed decision. We do not want the risk of getting a false positive for your campaign and data. All data is fake below, though, it gives you an ideas of what your spreadsheet could look like.
You may be asking yourself what each of the columns mean and how that might apply to your startup. Let’s runthrough of each columns in the spreadsheet and how it might apply to your startup.
Campaign: This is the name of your campaign. Try to name it something anyone new working on your marketing team will understand when they look at the campaign.
Type: What is the campaign about? Launching an ebook or new service, I put down what’s here. This way I can compare similar campaigns and see if I can spot any trends over the year.
Language: If you’re launching a campaign in a non-English market, I put it here. Similar to type above, this can help you start to spot trends in different markets. The subtle difference between countries across Europe is important if you are going to market across such a diverse region.
Spend: What you spend across ad platforms (Google, Facebook, Linked In ..ect). In the “Notes” column below I’ll say if one network drove more of the business. e.g. Maybe Facebook drove 90% of your ebook downloads for November.
Downloads: Launching an ebook or whitepaper tomorrow? This column tracks the downloads over the next 60+ days.
Blog Subscribers: People who subscribe to your blog or this could even be subscribers to your company newsletter if you don’t have a blog. If your startup has both, just add another column in your spreadsheet for tracking each. If your company auto-subscribes someone to the newsletter, if they provide an email for your blog. Please make a note of that.
Conversions (Con): This is your main KPI and usually is a customer signing up for your service or product. Making sure this is setup correctly in GA is important. Also, make sure that everyone runs through your customer sign up when they start at your startup. This is a great way to spot error and understand the process that your customers go through.
Last Non-Direct Click (LNDC): If you removed direct conversions in Google Analytics and looked at the second last click before someone converted and count that channel instead…that is LNDC in a nutshell. To grab this data, see the note towards the end of this blog post.
Assisted Conversions: This is about assisted conversions which is an excel SUM formula from the two purchase path columns. See the section below on assisted conversions and how to find the data in Google Analytics. A campaign that drives a lot of assisted conversions is neither good nor bad but a good thing to take into consideration when running campaigns.
Purchase Path: These two column tell me how many people took less than 5 clicks to convert VS over 5 click to convert on your main KPI. Understanding how long it takes the majority people to convert can help you understand what your customer journey looks like.
Assorted Campaign Details
Flight Dates: When did you launch the campaign. Always make a note if it spans multiple months as seasonality could come into play with your brand. Even the fact that some brands see a dip in sales in the summer when it’s nice outside vs rainy days. I see this a lot with ecommerce brands that sell clothes.
Month: Some campaigns cross over into multiple months. I put the month that the budget is getting assigned too. I try not to split campaigns budgets into multiple months as much as I can. The finance team at your company will appreciate you for this.
Region: Where in the world are you lunching this campaign. This goes nicely with the language column as it helps you spot trends in region data as you run more campaigns.
Notes: This is for campaign notes that may not fit anywhere else on here. Maybe you tweaked your audience targeting part way through the campaign, that would be a good thing to write down.
Why Does Purchase Path Length Matter?
If you look at the example below, not real data, you see that this business gets 90% of their business in 5 clicks or less. Knowing that the majority of people convert in less than 5 clicks means you can understand which of your campaigns is helping grow the business. There are three very logical and very good reasons for knowing the purchase path length of your business:
1) This helps set a benchmark to measure success against. Benchmarks are your friend! If a campaign had the majority of people convert at 10+ clicks, then that might be a campaign you do not want to run again.
2) I’d argue that if most customers take 5 clicks to convert and you’ve ran a campaign that has the majority of people converting under 5 clicks. Even if that campaign isn’t the last click before someone converts, the fact that people are converting means it’s helping move your business forward. This might be a campaign you want to put more money and resources into.
3) Any things over 5 clicks and especially when you see people taking 15 or 20+ clicks on their path to convert and your campaign is just one them. I’d argue these customers would have converted with any campaign they saw from your brand.
To find what your startups purchase path is, you need to login to Google Analytics. Head over to Conversions –> Multi-Channel Funnels –> Path Length and pick a 60 day date range. You’ll only see data if you’ve goals setup in Google Analytics. Over 60 days, how many clicks does it take until 80% – 90% convert with your business? This is the number you will use to measure assisted conversions against.
How To Find Assisted Conversions in Google Analytics
Login to Google Analytics and head over to Conversions –> Multi-Channel Funnels –> Top Conversion Paths. At the end of “Primary Dimension” you’ll see “Channel Groupings” with a little drop down arrow. Click that arrow and copy the MCF Channel Grouping template.
Now the new copy of the MCF Channel Grouping template should be available for you to edit, copy or share it. We only want to edit it right now. Click on the little pencil to edit your new channels grouping.
Editing this new MCF Channel Grouping will allow you to figure out, on a campaign by campaign basis, which campaigns are helping drive a conversion under or over your purchase path length of 5 clicks. First I renamed mines Paid Campaign – Multi-View but you can call it something that makes more sense in your startup. Click the pencil to edit “Paid Campaigns” and change the campaign name field to name of your campaign. Google should auto-fill this as you type. Click done and then save.
The last two steps are easy as you pick a date range for your report, I do mines week by week, and search for paid in the search box. This will bring up every campaign that helped drive a conversion for your company. Add up all the conversion under and over 5 clicks and fill out the spreadsheet.
Steps To Last Non-Direct Click (LNDC) Column
Head over to Conversions –> Attribution –> Model Comparison and pick Last Non-Direct Click from the drop down menu in the middle of your screen. Take the same steps above when you edited your MCF Channel Grouping template to look at each campaign on its own.
You don’t have to worry about remaking the MCF Channel Grouping template because it’s available in Google Analytics for any report that has channel grouping as an option. Now take the conversion number you get back on the paid channel under “MCF Channel Grouping” and that is what I’d put on your spreadsheet. Again look at what happens 60 days after each campaign launched. Again I do this on a week by week basis.
How To Analyze Your Startup Attribution Model
So you’re done right? Not even close because you’ve the data and now you need to do the hard work and analyze the data. You need to figure out where… Where do you spend more money? Where do you spend more time? …based on the spreadsheet below.
Some looking at the above sheet some might say spend more money and time on ebooks (campaigns 7 & 8) or spend even more on the conference (campaign 3), which are logical and solid answers. I ask myself first, can we scale the conference campaigns and grow it into a power house. The super long tail of conferences is no joking matter. Conferences are the long game and you shouldn’t join if you’re not willing to invest.
I’d look to put my money and time into product launches (campaign 9). The reason is the costs are low to start, there is some early traction across last-touch, assisted conversions and even LNDC. If people do buy better versions of themselves then we can build a great long-term campaign around a feature that helps customers become better versions of themselves. Outside of saving them time and money of course.
Every brand and company will be different and you’ll have to play with your worksheet and test out theories. However, for most brands, if you focus on widening those who know about you through ebooks & conference(s) and get people to convert because of your features & you make them rock. You’re on solid ground. That is it for this week. See you next time.