It’s that time of year again! After holiday work parties, family, food, more food, more drinks, even more food, and even more drinks…reality sets in. We’re all collectively over the hangover and ready for a fresh start.
Enter New Year’s resolutions, because you don’t want to be left saying “I’m in the same place I was last year, not again!” next year.
That’s how many companies feel about their customer data.
Customer data practices are a lot like having strong health habits. You need to know which behaviors to cultivate and have the discipline and willpower to regularly execute those actions towards your goal. Eventually, when done correctly and consistently, these behaviors become a habit. While, I’m the wrong guy to coach you into good eating and exercising habits, but I am former database consultant, a former head of marketing ops, and currently head of product at a new startup. I’m perfectly qualified to be your personal customer data trainer! My friends at Kapost invited me to share my top marketing and sales technology steps to optimize customer experience in B2B SaaS.
- Marketing Application Silo: Most tools used for execution have data
- Marketing Cloud*: Most execution is done at the sales automation or marketing automation level. However, most teams buy and implement tools in silos, not fully considering the customer.
- Customer Experience Operating (CX-OX): Each tool in the stack works together synergistically. Engineering efforts and home-grown features can also seamlessly be integrated. All customer data is collected and aggregated in a data warehouse.
*I would estimate that 80-90% of mid-market and 70%+ of enterprises fall into the marketing cloud category.
He then goes to say: “IDC predicts that 50% of all digital transformation efforts will fail due to the absence of the CX-OS.” Here is a diagram showing the categories of tools in a CX-OX.
The list of resolutions below is in order of architecture maturity. If you’re in earlier in your journey, focus on #1-5. If you’re farther along in your journey, look at #6-10. Without further ado, here is my top ten list of data resolutions. Go through these with your ops team and commit to making a change!
10 Ways to Improve Customer Data
1. We will Implement NPS & CES, then measure and improve on it obsessively.
It doesn’t have to be a long, complex survey—could be as simple as adding polling in your app, as a non-intrusive flyout, such as provided by Satismeter. Or it can be a something more comprehensive such as an email survey (although I find in-app easier to implement and more reliable). If you don’t have an app or your customer base is very complex, here is a good list of survey tools you can use. The Bottom line is if you’re not regularly checking your customer pulse, you’re flying blind.
2. We will enrich our CRM data and customer data comprehensively.
Many companies do a one-time data enrichment job, try a few vendors, get bad data, and quit. But, I promise you that if enrichment is done properly, it will lead to faster deal velocity and a better overall customer experience.
- Fill in whitespace at the Account, Contact, and Lead level.
- EVERY Account has a website.
- EVERY Contact has an email (preferably a business email)
- EVERY Lead has an email AND a website (preferably a business email)
- Use email address and website to records via a partner API. Firmographic and demographic fields are key for segmentation, lead routing, and analysis. How is your AE specializing in financial services supposed to get a financial services lead, quickly, without data enrichment?
- Test the inbound enrichment match rate against your vendor. If you sell to SMB the match rate is expected to be lower (<40%). The highest match rates I’ve found are from Clearbit. They have a killer Salesforce app too.
- Try different and sometimes use multiple data enrichment sources–not all vendors provide the same data signals and may have stronger match rates for you in certain areas.
3. We will eradicate bad/duplicate data and monitor our lead-to-account match rates.
Three of the most important database metrics are whitespace/field-completeness (covered above), duplication rate (how many emails/websites in your CRM are duplicate), and finally lead-to-account matching.
- Some vendors that specialize in this:
- Top picks for deduplicating: Ringlead, DemandTools
- Top picks for lead-to-account matching: Engagio & Leandata
- Why is this essential for customer experience? It all has to do preventing fragmenting between buyers at the same company and also faster follow-up times from sales/ customer service:
4. Create manageable, consistent business segments in all systems
A sales leader at Salesforce.com once told me: “We have a policy here called, stay in your lane.” I didn’t grasp how important that concept is at the time. Now, it’s crystal clear why it’s critical to a good customer experience. Regardless of how your business and sales territories are divided, make sure that you enforce this division and agree on it across the business. Make sure ALL your systems and internal users and can identify and enforce these segments. Here are some questions you can look in the mirror and ask to check if you employ “stay in your lane”:
- Does your app or product represent ‘know’ the important things that Salesforce knows (and vice versa)? Hint: If no, this is because your product database doesn’t sync tightly with your CRM
- Is person logging into your tool from a named account?
- Can I tell the last time somebody logged into my tool from Salesforce?
- Do named-account prospects ever end up on the phone with a Jr. SDR? Hint: this is due to poor lead-to-account matching and routing.
- Is it easy to create “Mutually Exclusive, Collectively Exhaustive (MECE) segments on buy-stage in my marketing automation tool? Hint: If no, remove complexity and fill in white space.
A deck of cards is a perfect MECE example. Every card has exactly one rank (13 values) and suit (4 values). If you want to create a segment, it’s easy. Similarly, every person in your database needs the right company and buying stage assigned.
5. Create ideal and non-ideal customer profiles
Many companies on the market have tools that can help you create ICPs. Companies like 6Sense, Everstring, Mintigo, and Infer specialize in Predictive Analytics and identifying Ideal Customers. They are very cool. The main downsides are 1) they are expensive, 2) they rely on historical data, 3) they take several months to roll out and be adopted by sales. Another approach is to create simpler ICPs based on enriched CRM data. For example, if you were using Clearbit alone, you’d have access to an over 100 company and person signals with which to create ICPs. I could for example quickly see how many of my customers use Marketo → if 90% of my revenue comes from Marketo customers, then my ICP includes “has Marketo.”
ICPs not only help lead routing and scoring — but they give you clues as to where to create content. In the example below more than 80% of revenue is coming from just six industries! I would proportionally invest in content for these six industries.
Here is a simple chart that looks at revenue and number of customers by industry.
6. We will formalize and adopt personas across your organization.
|Personas aim to get at specific psychographic motivations of your target buyers. If your ICP is Digital Marketing Managers at Mid-Market SaaS companies → your aim should be to generate content that would be useful to different categories of this ICP.
Profiles help you identify personas.
Personas answer the “what’s in it for me?”
If segments are the canvas, and profiles are colors, you can think of personas as the shades of each color.
How to develop personas:
- Interview customers (at least 20 to start) about why they bought your solution (this is called top-down persona development).
- Use ICPs to deduce who your personas are. If you mostly sell to Enterprise IT Directors, then categorize what Enterprise IT Directors care about. Tip: more than 5-6 personas becomes very hard to manage, and can spread your efforts too thinly
- If you only have wall art personas, that’s OK. It’s a good start. Socialize them, get feedback.
7. We will uncover and fill our content gaps and product gaps to build a unified strategy around your persona.
Gap analyses compare what you have today vs. what you should have. But the goal should be to create a heat map to see what’s missing in your strategy. One one axis you put the targets (persona, industries, buying stage) and on the other access you put content coverage (counts or an estimated coverage score).
- Manual – this can be done by having your content team individually count the sum of documents relevant to each buying stage, persona, etc.
- Automated – using Natural Language Processing and Machine Learning technology is rapidly
This heat map from Trulia shows when users do searches in their app.
8. We will *finally* create a single view of the customer by consolidating 1st, 2nd & 3rd party customer data.
Once your core data silos are clean, complete, standardized and tightly integrated with each other. It’s time to create a customer data platform! Decide on your requirements for your data project based on the 5 V’s (Volume, Velocity, Variety, Veracity, and Value). You’ll encounter three major Varieties of data:
Structured / Relational data:
- Stored in rows and columns (CSV, Excel, Salesforce.com, Marketo, most SaaS)
- Records directly related to each other
- Each record has a finite set of attributes.
- They are eligible to be a prospect (Y/N)
- A company has a domain, e.g. “company.com”
- A related person has an email address, e.g. “firstname.lastname@example.org”
- That person has a title, e.g. “Head of Product and Stuff”
- Has a predefined order and loose rules
- Records look similar but can vary greatly in size and texture
- Has no predefined order
- According to Wikipedia, 80-90% of all data is unstructured
- Videos, recordings, pictures and random snippets of text are all examples
Collecting a variety of voracious data is the fastest way to create rich dimensionality in your customer database. You’ll want to figure out a solution for each different type data variety. Have a strategy in place whether to build or buy a CDP. This isn’t easy, but you’re in luck, CDPs are becoming cheaper and more approachable to set up due to the great integration middleware available, like segment.com.
9. We will operationalize our personas.
Sales, marketing, product…all playing from the same playbook. This is the missing link! I took the below iPhone pic from this year’s SiriusDecisions tech exchange in Austin, TX. Kerry Cunningham who is the lead analyst covering predictive analytics gets really excited when talking about this “next” category.
Think back to the iceberg above, what if we could predict what is below the iceberg? Lead routing and tactic matching based on the “What’s in it for me” and psychographics. How can we digitize putting the customer first?
10. We will use machines to solve machine problems and use humans to solve people problems.
- Automated content audits
- AI for calendaring
- Real-time web personalization
- APIs, APIs, APIs, APIs
- Predictive Email
- Predictive Sales Plays
The only way to truly become/create a CS-OX — to make every interaction a personalized and meaningful one, with every customer — will be to rely on machines. Every time you hear or say the word “manual process” at your company, ask “how can I automate this?” Ask, “how can I leverage data to create a better customer experience?.
But at the end of the day be resolute towards your customers. Don’t forget to add a uniquely human and personal touch whenever we can. Handwritten cards and chocolate will never go out of style.