I have spent 25 years doing data analysis, a skill by which long ago I eliminated from my resume, not wanting to become a victim of my talent, yet it definitely comes in handy.
You cannot make data driven decisions if you don’t have good data and data you can trust. However, sometimes data can be misleading or worse, wrong.
The most important thing you can do as a CMO is make sure that you have a handle on your data when you start at an organization. No matter what, it’s the single most important thing you can do because it will help you understand the state of the business, what needs your focus and attention first, and assess your talent.
Good data comes from enforced data hygiene
Determine the data you want to track and then enforce that within the systems. Most CRMs will allow your System Administrator to make fields mandatory to be filled out by Sales or block fields from being updated by anyone or have fields auto-updated from third party sources. You may still have some fields that need to be manually updated, but you can also outsource that or hire a very cheap intern to do it.
For example, BuiltWith allows you to see a company’s tech stack based on tags it sees/scrapes from their site. You can actually pipe that into your CRM so that Marketing or Sales can use it to better tailor their messaging or to create competitive lists to target.
Or another example is when a sales rep closes an opportunity, as a CMO you want to know who were you competing against, why did they buy/not buy your product/service (win/loss reason), and more. These are all fields that can be made mandatory so that the rep cannot close out the opp without filling these fields out. Of course it’s like filling out a survey, you will need options like “did nothing”, or “no competitor” or whatever is most appropriate for your business.
MQL Data
The biggest problem I see CMOs and marketers struggle with is the MQL data. How do you know where it came from? Who gets credit for it? When did it MQL exactly? When did it move to SAL or SQL? All of these questions can be answered by inputting custom fields into your CRM.
For example, I’ve always used Campaigns in Salesforce to track MQLs so I can see what channel and campaign is bringing in MQLs. I also always create a “Date became MQL” custom date field that auto timestamps when something becomes an MQL.
And while campaigns are great, some companies don’t use campaigns, so then what? You can still create MQL Source to populate from your UTMs or MQL detail to populate based on referrer URL. So there are ways to get to the granularity of this without needing an attribution tool.
What if the data is non-existent or sh*t?
When there is no data (good or bad), or there is bad data, your best bet is to take the opportunity to create new definitions of what you want to track and what’s important to the business and write it down to socialize with the entire company. (MQL = x, SQL = x, Pipeline = x, etc).
You’ll need to get agreement and buy-in from Sales, CEO, and Finance on what SQL is and what pipeline is because it most likely will impact SDR/BDR comp plans. Plus it sets the stage for when the inevitable conversation of “these leads suck” comes up. In companies I have worked at in B2B SaaS, I’ve always defined Pipeline as opportunities that had $ associated to them and I always worked with SalesOps to make it a mandatory requirement in the CRM for Sales reps to have to put in $ amount when moving a deal to Stage 2. After all, a $0 opportunity is a pipe dream, not pipeline.
The method I’ve always followed for opportunities has been:
Stage 0 = meeting booked
Stage 1 = meeting held (usually Discovery)
Stage 2 = pipeline created
This helps you compensate BDR/SDRs for meetings booked or for qualified meetings held (because if it doesn’t move to Stage 2 after the first meeting then it’s not a qualified meeting and therefore shouldn’t be compensated).
While this seems rudimentary and simple, you’d be surprised at how many MOPS people over complicate this process. It shouldn’t be complicated. Complicated doesn’t equal sophistication or smart & savvy, it equals stupidity.
Most importantly, do not create your data architecture around corner case scenarios. Use the 80/20 rule here. Otherwise, you will wind up with a mess.
What if the data is solid, but doesn’t paint a good picture?
Data analysis is about storytelling. You’re a marketer, you should be great at this already. So ask yourself, why isn’t the data painting a good picture? What story do you want to tell? Just because dashboards exist at a company when you walk in doesn’t mean they are right or supportive of the strategy you are going to drive. I’ve walked into plenty of organizations where the data looked accurate or looked good, only to dive into it and realize, it tells a nice story, but it’s not accurate.
As an example, suppose you walk into an organization and the Marketing team is crushing it on their MQL/MQA goals, but the Sales team is not crushing it on their quota attainment. That usually means one of two things, either: 1) the ASP/ACV is too low so the reps need to close more deals to hit attainment, or 2) the quality of the leads are not good or not as good as they could be.
To put it better into context, suppose you are going up market and targeting Enterprise, but the MQLs you are bringing in are mostly SMB or lower end MidMarket, so therefore the ACV/ASP is getting driven down. Now you might say, well, then they are not qualified and we shouldn’t have passed them to Sales in the first place. Sure, that might be true pre-Covid when you could be more picky about who you sold to, but that’s not reality in 2024.
You are the story teller and you need to tie it back to the overall company objectives. Design simply, you can always back and tweak your data. Which brings me to my last point.
How far back should you go to fix data?
I love this question because its where you can make yourself miserable if not thought out properly. You do not need to go back to the beginning of time to update data. You need to go back a minimum of 2 quarters and at most 4 quarters to correct data if it’s worth correcting so that you have QoQ or YoY comparisons. But, it really depends on what you are fixing. Sometimes, you don’t have to go back and fix it at all.
What do I mean by that? Say for example you change the way you score MQLs. Maybe you only counted hand-raisers as MQLs previously and now you’re introducing scoring and counting folks who have downloaded content. Should you go back and score people prior to the change? The real questions to ask are: will this impact my current quarter KPIs? If yes, then score them. If not, then don’t bother.
When it comes to deciding whether to fix data or not, I ask these questions:
Will the board/CEO/Finance care about this data?
Does it impact my current quarter KPIs? or impact anyone’s current quarter compensation?
Will it impact QoQ or YoY metrics down the road? (and will anyone really care?)
What will fixing this data get me? If it provides further information to make a larger more informed decision, then it should be weighed against the amount of effort it would require.
If I determine I need to fix it, how far back should I go? 1–4 quarters should be the max.
Key Takeaways
Data is your friend and the sooner you get a handle on it, the easier your job becomes. Don’t over complicate it, start out simple with the most basic things you, your CEO, and the board care about and build from there. Remember, you are the storyteller, use the data to tell your story.
If you liked this article, then you may like: