Don't hire a marketing expert. This is a bad first step. You should define a very specific subset of customers and talk to them. Find out their problems after talking to 10 very targeted people that fit into a user profile. I recommend reading "The Mom Test" (it's a book about how to phrase customer development interview questions) as a basis for how you talk to potential customers.
User acquisition efforts are not only measurable on a larger scale. You may have this idea because the people you worked with may have had a big company background. For a startup, it is very necessary for your advertising campaign to define a cost per click, find a conversion rate, and then test many different advertising phrases and avenues. This helps put a direct cost per user acquisition price on each campaign and you can compare apples to apples. It takes "voodoo money burning rituals" and turns it into numbers that are measurable.
If you rely on free marketing channels without a clear focus on who is using your product, it is very likely that you have 10 customers who fit 10 completely different segments with 10 different ideas of which features will give them value.
They key is learning. When you estimate future software releases, you typically point a story using things you have learned from previous, similar stories. When you attempt a new advertising campaign, you should use past learnings to determine which ads and messaging will work for different segments - how much they will cost, and how many users you can expect to get based on past results.
In summation, startup marketing should have a focused, scientific, numbers driven, methodical approach - just like writing software. If your marketers are not doing that, you probably aren't working with the right marketers.