Data science and analytics

Answers to specific business questions from the data you already have, with an honest account of how certain they are.

Typical questions

Each engagement starts from a question that someone has to act on.

Why a number moved: Untangling a fall in sales or traffic when several things changed in the same few weeks.
Whether a change worked: Measuring the effect of a launch, a price change or a campaign, even when no proper test was run.
Forecasting: Demand, revenue or capacity, given as a range with the odds attached.
Customers: Which customers stay, which leave, and what each group has in common.
Unstructured data: Reviews, support tickets and documents turned into something that can be counted.
Checking the figures: Whether a dataset, a report or a vendor's claim survives a second look.
Dashboards people use: A few measures everyone agrees on, built from data they trust.

When several things change at once

A drop in traffic or sales rarely has one cause. A search engine update, a site redesign and a new competitor can all land in the same three weeks, and a report that blames just one of them is often guessing.

The methods that help come from research. One is to compare groups that differ in a single way. Another is to find the natural experiments a business runs without meaning to. Sometimes the honest answer is that the data cannot separate the causes, which is worth knowing before money is spent on a guess.

How to start

The brief

Write down the question and the decision that hangs on it. Say what data you have, even if it is messy.

Your data

The analysis runs on your data where it lives, or on a copy that you control.

Who does the work

Amos Weiskopf founded Tevelis. He built the text analytics tools that a Fortune 500 company's marketing, R&D and strategy teams worked from, and he writes the code for his own analysis.

Questions buyers ask

What is the difference between data science and analytics?

Analytics describes what happened, through reports and dashboards. Data science uses statistics and models to explain why it happened, or to predict what comes next.

Do we need a data scientist on staff?

Not always. If the questions come up a few times a year, outside help usually costs less. If they come up every week, it is worth hiring someone.

Our data is messy. Is that a problem?

Most data is. Part of the work is finding out which parts can be trusted, and the answer says so.

Can you tell us what caused a drop?

Sometimes clearly, sometimes only in part. When the data cannot separate two causes, the report says that plainly.

How long does an analysis take?

The timetable depends on the question and the state of the data, and is agreed in the brief.

What does data science consulting cost?

That depends on the question and how much data work it needs. The price is agreed after the brief.

Engagements by written brief.

Work begins from a documented problem statement, not a sales call.

Write a brief Or book a meeting