Signs.com planned its growth on a 30% repeat-customer rate from its analytics, and its transaction records showed 8%. The customers who did come back were businesses, so acquisition moved toward accounts. Reviews on the best-selling product lifted its conversion more than 10%.

The result
30% reported, 8% real
Repeat customers in analytics versus transaction records. Reviews on the best-selling product then lifted its conversion more than 10%.
Signs.com believed 30% of its customers came back to buy again. The number had sat in its reports for years, and the growth plan rested on it. If almost a third of buyers returned, losing a little on a first order would pay off on the second.
So the team built for people buying signs for personal projects and gifts. It found new vendors, designed new landing pages, researched competitor pricing and launched about 20 new products for that buyer. The work took most of the marketing team's year. Those products added almost no new revenue, and in December 2017 the company posted a loss for the month.
The 30% came from Google Analytics, which counted visitors who returned within a rolling 30-day window. Purchase history showed something else: across all time, 8% of customers had bought a second time. Analytics was over-reporting retention by almost four times, and every forecast built on it carried the error.
That 8% mattered more than its size suggested. It produced about 80% of revenue and all of the profit. So we picked up the phone and called them. The repeat buyers were businesses. When we called accounts that had stopped ordering, the person who set up the account had often left the company, and the new hire went to a different vendor.
The company had a service that fit those buyers: an ordering portal where a business could upload its artwork once and have it fitted and shipped to each location. Few customers knew it existed.
The fix had to show the sales team which accounts to call before they lapsed, using each account's own buying cycle instead of a company-wide average. It had to tell marketing which kinds of businesses were worth acquiring. And the quick wins had to come from pages that already carried revenue.
With a developer, we built an attrition score for every customer, based on how often that customer usually ordered. An account that normally bought every 30 days and had gone 45 scored 1.5, and the higher the score, the higher the risk. It lived in the company's admin system, where anyone could see it.
We classified every customer by industry code and found which industries were the most profitable, with franchise businesses at the top. Acquisition moved from individual buyers to accounts. The website became a source of business leads, and business development reps built each account, brought it onto the ordering portal, and kept two to three contacts per account so one person leaving did not end the relationship.
We added reviews to the best-selling product, which made up about 10% of revenue on its own. And when sale weeks pushed a customer service team of about 30 people into 60- to 80-hour weeks, we split each sale into two three-day windows two weeks apart.
Signs.com kept the attrition score in its admin system, along with an account-based acquisition plan built around its most profitable industries and the ordering portal. The company's business development rep used the score to decide which accounts to call first.
Conversion on the best-selling product rose more than 10% after the reviews went live, on a product worth about 10% of revenue.
Splitting the sale windows ended the overtime and the catered late nights while revenue and profit both went up, because customers got answers faster.
Key takeaway
When your analytics and your transaction records disagree, trust the transactions. Then call the customers who came back and ask them why.
Industries: B2B
Find problem areas today with the Market Health Analysis. It's $350, credited toward your Diagnostic.
Get your Market Health Analysis