Why contact data goes stale
A list of business contacts is correct on the day it is made. From then on, the world moves and the list does not.
People change jobs. They move to another company, and their old work address is closed or forwarded to someone else.
People change roles. A buyer becomes head of purchasing; the title in the list is now wrong.
Companies change. They move offices, merge, rename themselves, switch to a new domain or close down.
Details change. Phone numbers are reassigned, email formats change, people change their names.
How fast this happens differs a lot by industry, country and role, and there is no reliable public figure for it. This worksheet therefore uses invented monthly rates. The method is what matters: it works with any rate you measure in your own data.
Under the GDPR, personal data must be accurate and, where necessary, kept up to date (Article 5(1)(d)). An out-of-date contact list is not only a sales problem.
Task 1: Which fields age fastest?
Sort these fields of a contact record from "changes most often" to "changes least often", and give a reason for the top and the bottom one: company website, job title, direct phone number, work email address, company postal address.
Solution to task 1
A sensible order, fastest first: job title, work email address, direct phone number, company postal address, company website.
The job title changes with every promotion and every new job. The email address changes when the person leaves the company. The company website and domain belong to the company, not the person, and change only when the company renames, merges or closes. Other orders can be right if the reasons are good: in some industries phone numbers change faster than emails.
Task 2: One month, two months
Northwind Traders has 1,000 correct contacts on 1 January. Each month, 2 percent of the contacts that are still correct become wrong.
a) How many are still correct after one month? b) How many after two months? c) A colleague says "after two months, 960". Why is that slightly wrong?
Solution to task 2
a) 1,000 × 0.98 = 980. b) 980 × 0.98 = 960.4, so about 960. c) The 2 percent in the second month applies to the 980 contacts that are still correct, not to the original 1,000. So 19.6 contacts become wrong in month 2, not 20. The difference looks tiny now but grows every month.
Task 3: Compound decay over two years
Because each month takes 2 percent of what is left, the number of correct contacts after m months is:
correct after m months = 1,000 × 0.98 ^ mFill in the table. Round to whole contacts.
Months | Correct contacts | Wrong contacts | Share still correct |
|---|---|---|---|
6 | |||
12 | |||
24 |
Solution to task 3
Months | Correct contacts | Wrong contacts | Share still correct |
|---|---|---|---|
6 | 886 (885.84) | 114 | 88.6 percent |
12 | 785 (784.72) | 215 | 78.5 percent |
24 | 616 (615.78) | 384 | 61.6 percent |
After one year, more than one contact in five is wrong. After two years, almost two in five.
In a spreadsheet: months 0 to 24 in A2:A26, and in B2 the formula =1000*0.98^A2, copied down.
Task 4: The linear shortcut
Someone estimates: "2 percent a month is 24 percent a year, so after 12 months 760 contacts are correct, and after 24 months 520."
a) Compare these numbers with your results from task 3. b) Explain in one sentence why the shortcut is wrong. c) What would the linear shortcut predict after 50 months, and why is that impossible?
Solution to task 4
a) Linear: 760 after 12 months and 520 after 24. Compound: 785 and 616. The shortcut is too pessimistic, and the gap grows with time. b) The 2 percent applies only to contacts that are still correct, and there are fewer of them every month, so fewer become wrong each month. c) 50 × 2 percent = 100 percent, so zero correct contacts. With compound decay, 1,000 × 0.98 ^ 50 = about 364 contacts are still correct. A share can never fall below zero.
Task 5: Fields at different speeds
Wide World Importers measured (in this invented example) how often each field of its 1,000 contacts becomes wrong per month. Calculate how many of each field are still correct after 12 months, rounded to whole contacts.
Field | Becomes wrong per month | Still correct after 12 months |
|---|---|---|
Job title | 3 percent | |
Work email address | 2 percent | |
Direct phone number | 1.5 percent | |
Company postal address | 0.5 percent | |
Company website | 0.2 percent |
Which two fields should the team re-check first?
Solution to task 5
Field | Becomes wrong per month | Still correct after 12 months |
|---|---|---|
Job title | 3 percent | 694 (1,000 × 0.97 ^ 12 = 693.84) |
Work email address | 2 percent | 785 (784.72) |
Direct phone number | 1.5 percent | 834 (834.13) |
Company postal address | 0.5 percent | 942 (941.62) |
Company website | 0.2 percent | 976 (976.26) |
Re-check the job title and the work email address first. A wrong title makes the email miss the right person; a wrong address bounces or reaches someone else.
Task 6: A re-check schedule
The sales team wants at least 90 percent of its email addresses to be correct at all times. Email addresses go wrong at 2 percent a month.
a) After how many whole months does the share of correct addresses fall below 90 percent? b) How often should the team re-check the addresses? c) Job titles go wrong at 3 percent a month. How often must they be re-checked for the same 90 percent target?
Solution to task 6
a) 0.98 ^ 5 = 0.9039 (90.4 percent) and 0.98 ^ 6 = 0.8858 (88.6 percent). The share falls below 90 percent in the sixth month. b) At least every 5 months. A schedule of every 4 months leaves a safety margin. c) 0.97 ^ 3 = 0.9127 and 0.97 ^ 4 = 0.8853, so job titles need a re-check every 3 months.
Bachelor level: the exact limit is the logarithm ln(0.9) ÷ ln(0.98) = 5.2 months for email and ln(0.9) ÷ ln(0.97) = 3.5 months for job titles.
Task 7: What decay does to bounces
Northwind Traders emails its whole list of 1,000 contacts after 12 months without any re-check. Assume (an invented figure) that 60 percent of the contacts that became wrong produce a hard bounce, because the old mailbox was closed. The other 40 percent still accept the email, but the person has left.
a) How many hard bounces should Northwind expect? b) What is the bounce rate of the send? c) What happens to the other 40 percent of wrong contacts, and why is that a problem too?
Solution to task 7
a) Wrong contacts after 12 months: 1,000 − 784.72 = 215.28. Hard bounces: 0.6 × 215.28 = 129.2, so about 129. b) 129 ÷ 1,000 = about 12.9 percent. Mailbox providers read a bounce rate like that as a sign of a list that is not looked after, which can damage the sender's reputation (see the guide to sender reputation). c) About 86 emails reach a mailbox that someone else now reads, or nobody reads at all. The message misses its person, and personal data about someone who left is still being used, which conflicts with the GDPR principle of accuracy.
Task 8 (harder): The half-life of a list
After how many months is only half of the 1,000 contacts still correct, at 2 percent a month? Find it by trying values in a spreadsheet, or with logarithms.
Solution to task 8
0.98 ^ 34 = 0.5031 and 0.98 ^ 35 = 0.4931. After 35 months fewer than half of the contacts are correct. With logarithms: ln(0.5) ÷ ln(0.98) = 34.3 months, just under three years.
For discussion
Your company finds a list from three years ago on a shared drive. What do you do with it?
How could a team measure its own decay rate instead of guessing? (For example: count bounces and "no longer works here" replies per month and divide by the number of contacts.)