-
Optimise your marketing: stop guessing, start uncovering insights
Last stage of the marketing flywheel, and my specialist subject if I ever went on Mastermind.
Most optimisation advice starts with “test more things”. But I urge you not to jump straight to solutions. Figure out where the biggest problems are first.
The graphic below shows the research toolkit. It runs from quantitative data (what’s happening) to qualitative data (why it’s happening). Analytics and heatmaps show you where people drop off. Surveys, sales interviews and chat logs can tell you what (almost) stopped them from buying.
The good news: you don’t need all seven. For most founder-led businesses, three will do for now.
- Analytics, to find where you’re losing people and the size of the problem.
- Heatmaps and session recordings, to watch how users interact currently.
- Your last ten sales conversations, to hear the objections in your customers’ own words.
That last one is free and sitting in your inbox, but often gets ignored.
One change based on evidence beats five changes based on hunches. Data and insights come first, then decide the approach, then make the change, then measure. Rinse and repeat.
Which of these have you actually used? Drop it in the comments.
New Office Hours time: Wednesday 1pm if you want to discuss how to optimise your marketing efforts.
meet.google.com/qhk-qowa-yvxSandip
Emma Selby *Flexspace Operator*, Stuart Morrison and 2 others2 Comments-
@Sandip – having been CRO focused in our business at one point, the biggest issue we found was smaller operations have non-statistically significant data. How do you overcome that. Sure, there’s the sense-check, and you get an idea over time what is likely but how do small business overcome the trap that we saw thinking because you had ten new enquiries after doing some work that the work = uplift, when it may have been just a trivial statistical variance, due to indeterminate cause?
10 more people clicking a button when the visitor numbers are 1-2K and variance is +/- 500 people a month, hard to determine if the work you did was the difference or just “one of those things”?
-
@stu You’re absolutely right, A/B testing isn’t appropriate for many businesses – you need tonnes of traffic and conversions in the first place in order to have any hope of a statistically significant result. However, A/B testing is just one method of helping make better decisions – sure, it’s the gold standard (if done right) but there is a whole toolkit of methodologies you could use to glean insights on your customers and how they behave. You can still test different designs etc, but it would have to be a linear test – i.e. review the baseline, make a change based on insights gleaned, see the impact on the metric. It’s not as reliable as A/B testing as it’s done in different time periods, and many factors could impact the result like seasonality etc – however, it’s a decent fall back position. The key thing is having a solid hypothesis based on actual customer insights.
1
-
