Even without a crystal ball, I can confidently predict that much of the marketing data you use for your marketing campaigns is rotten. In my corporate career at IBM it was a constant area of frustration, as it has been in almost every client I’ve worked with over the last 4 years, irrespective of size.
According to ZoomInfo around 30% of marketing professionals change roles every year, and 60% of us change job titles. Furthermore the data you have on the contacts and their associated companies is almost certainly incomplete and inconsistent. It’s not uncommon to see 50% of a company’s database records without job title, industry or number of employees – the basis for even the most straightforward targeted campaign. Consequently vast swathes of your database is never contacted, or are included in campaigns that are totally inappropriate for them. If you can generate sufficient activity from inbound marketing, of course, that may not be a problem; but for most companies outbound marketing is still essential, and therefore so is decent data. Poor data quality hits in-bound performance too e.g. if you start nurturing contacts with
incorrect contact details.
Even in a tiny one-man business like Purple Salix the same challenges arise. I use a simple CRM system (the free HubSpot CRM) to manage my business, and although my revenue tends to come via recommendations rather than extensive outbound email marketing, being able to segment my prospects and targets is an essential element of prioritising my efforts. Consequently I spend a considerable amount of time trying to improve the return and insight I can get from my CRM data. Even as a one man band my CRM currently has around 2,500 contacts across 1300 companies!
Sorting the Good from the Rotten
There are essentially 4 steps involved in moving forward:
- Understand what data is still valid
- Correct the data that is inaccurate
- Fill in the gaps that still remain
- Add additional data that can add additional insights
Historically the approach to this would have been to have a telemarketing agency attempt to contact the individuals in the database, or to call into the company switchboard to try to uncover key decision makers. But this approach is extremely labour intensive, somewhat limited in what it can achieve, and ultimately time-consuming and expensive.
For a business of my size, this really wasn’t a viable option. However, much of the data we would need to clean already exists within the internet, and much of it in the public domain.
Probably the best source of accurate contact-level data is LinkedIn, since it is in the interest of LinkedIn members to keep their key job-role and company information up to date.
Indeed within my own primitive data maintenance process, I regularly download all my first-level LinkedIn connections. This provides me with some basic information on my connections (name, company, position & email address). Comparing the latest LinkedIn records with the existing CRM records is a great pragmatic way of identifying which CRM contacts have moved on to different roles within the company, or have moved on to a different company altogether. But while this approach has value for me as a singleton, it is not a scalable solution since I can only export LinkedIn details for my 1st Level connections.
Another challenge with this approach is that the email address connected to the LinkedIn account is as likely to be a personal email address as a business one. And, of course, don’t confuse having an email address with having the owner’s permission to use it within email campaigns. But as I said earlier, I don’t really do email campaigns very much…
So all in all, a useful approach for an individual, but not scalable for business purposes.
Sounds like a Job for Contact Intel!
I was therefore very excited to learn of a new offering called Contact Intel from UK-based Cyance. Contact Intel takes as similar approach to my manual LinkedIn approach, but combines it with other publicly available social data to allow a potential matching against 650 million global, self-updating records. Additionally it can augment your data with additional fields at both the contact and company level such as firmographic variables and contacts’ skills, interests and groups to help improve targeting and messaging.
Moreover Contact Intel can find new contacts that you are not connected to that you would like to be and provide you with the same information.
All of this can then be used to validate what you have,fill in the gaps in your data and append data on new prospects.
A little taster, Sir?
Before I would feel confident to suggest that my clients should consider leveraging Contact Intel within their own environments, I wanted to ‘kick the tyres’ a little myself. So I handed over an extract of my CRM data to see what insights it could provide. In the interests of full disclosure, let me declare that I have now agreed to become a partner for this solution, and so will receive a small commission for any of my clients who use this service.
Having handed my file over, I got an updated file returned just a few days later. I found the results fascinating:
- 44% of my CRM Records found a match against the Contact Intel data. Clearly the level of matching will be entirely dependent upon the source data – actually I was pretty delighted that the software was able match almost half of my data – particularly given the amount of my database that only had personal emails to work from.
- For all of these records I now have additional consistent information that is great for targeting purposes – LinkedIn profiles summaries, location, Skills, Interests, LinkedIn groups they subscribe to, etc. And at the company level it has helped me populate mailing address fields, switchboard numbers, web address, Industry/SIC data, number of employees, revenue and more
- For 25% of my records it reported a different company on Contact Intel. On the face of it this implies that 25% of contacts have moved on to a different company; however, on further investigation, while this was usually the case, in some instances the contact hadn’t actually moved, but they had added an additional role on LinkedIn (e.g. a volunteer role). So a manual scan was needed to highlight those discrepancies.
- An additional service (that I didn’t use in this experiment) could also have helped identify, in the case where someone had changed roles, who was now the individual occupying that seat!
- 11% of emails were updated. This is really a side-benefit rather than the core purpose of the exercise. However, it does suggest that removing the old email addresses would improve any deliverability scores and avoid wasted effort.
- Job Title was updated in 17% of contacts. Of course that’s hugely valuable if you are targeting particular job roles or job levels.
- Re-importing the data back into my CRM took a little thought as it needed me to add a number of new fields and consider carefully which fields should be added and which would be safe to overwrite. In my case, I found the most effective approach was to manipulate the new and historic data within a spreadsheet before re-importing back into HubSpot as separate Contact and Company files, but it will depend on your particular scenario and Cyance can help with this if needed.
Shall I Put that in a (Purple) Box to Take Away?
Most clients I speak to have serious issues in the quality of their marketing data. There are a number of reasons why marketing leaders need to finally grasp this nettle:
- Ignoring the problem won’t make it go
away. In fact it will make the problem worse.
- If you haven’t done an audit of your data recently (at least in the last 6 months) you will inevitably have a significant and unusable amount of data within your database. Just think
about the number of “xxx has got a new job” notifications you’ve had from LinkedIn this year.
- If your company has merged or acquired other businesses the problem of rationalising two data sets makes matters much worse.
- If you are considering moving to a Marketing Automation platform, it’s important to clean your data first – most vendors price their licenses on the basis of the number of contacts you have. There’s no sense in paying over the odds for contacts that you could never use.
So if you’ve been struggling to figure out how to go about getting your data into a more healthy state quickly – let’s have a chat. This approach may not be suitable for everyone, but you absolutely need to put in place a regular ongoing data maintenance approach. If you have a large database, the smart thing might be to test the approach with a relatively small sample of a few thousand records. Based on my experience with my data, it can deliver significant actionable returns for only a small financial investment.