Business level strategy articles summary
Ella van de Loo
Week 2
Boyd, J. L. and Bresser, R. K. 2008. Performance implications of delayed competitive responses:
evidence from the U.S. retail industry. Strategic Management Journal, 29: 1077-1096
Abstract
The timing of competitive actions and responses is a key management concern that has important
performance consequences. This study focuses on the timing and consequences of competitive
responses. Theory predicts a negative linear relationship between response delay and responder
performance mirrored by an opposing positive linear relationship between response delay and first
mover performance. In contrast, our study suggests that response delay has a curvilinear relationship
with responder performance, and a linear relationship with first mover performance. We test our
propositions using retail industry data and discuss the implications.
Theoretical first mover research in economics, marketing, and management advocates quick
responses to competitive challenges to limit the first mover’s ability to build competitive adv.
- Economics: highlights the profit advantages of imitations -> ‘fast second strategy’ is
considered more profitable than acting first. Second may enjoy cost advs because it can learn
from rivals’ mistakes and improve upon the original idea.
- Marketing: recommends aggressive response – fast marketing mix reactions to entry such as
price or advertising adjustments are advocated to prevent the first mover from building
major market share and profitability advs.
- Management: importance of aggressive and fast response as well – to prevent first mover
from building barriers such as buyer switching costs and to signal incumbent (=zittend)
commitment to defend market positions.
General rule: quick and vigorous retaliation (=krachtige wraak) is necessary to limit an attack.
Theoretical research on response timing argues: attacked incumbents need to respond as quickly as
possible, because response performance will be increasingly limited as a first mover becomes more
established suggesting a neg linear relationship between response delay and responder
performance.
Response delay = the amount of time it took a firm to respond
Response order = the responder’s position in a temporal series of responses.
The paper focusses on theorizing on dynamic, competitive environments as has been the norm in
previous studies. In such industries, firms are motivated to compete aggressively because industry
structural conditions offer little protection for market positions.
A response timing decision is a balancing act, and its effectiveness depends on a firm’s ability to
assess market dynamics correctly and develop an effective response. A balance has to be found
between the risks of premature entry against missed opportunity. If a firm responds too early
because it has not waited long enough for uncertainties to subside or to develop an effective
response, its performance is likely to suffer. If a firm responds too late, it may not be able to expand
successfully because earlier movers will have gained enough leadership to dominate the market.
Thus, it might be a curvilinear relationship: on average, there will be lower success for fast
responders, higher success for responders with intermediate delays, and lower success for late
,responders. Responders with intermediate delays can be expected to outperform fast and later
responders because a measured delay will allow for designing an appropriate response without
moving too late.
Firms respond to competitive challenges too quickly or too slowly for several reasons: bad luck (when
a competitor makes an appropriate response timing decision based on available info at the time of
response, but unforeseen developments invalidate this analysis), conscious choices (based on correct
analysis of the market dynamics, but the responder consciously chooses to deviate from the
appropriate response timing scenario and accepts those risks), blind spots in competitor analysis
(response timing errors occur because a firm’s response timing is based on an complete/incorrect
analysis of the available info).
Sometimes firms consciously choose to respond fast although they know that a slower response
might be more advantageous. Exp: when important stakeholders expect fast responses to
competitive moves such as a price cut, attacked incumbent may respond fast by for exp also reducing
prices to maintain customer loyalty.
Firms may also choose to delay a response although they know that a faster response might be called
for. Their hope is that a delay will not hurt their position. Exp: firms may realize that they need to
obtain new resources before they can respond effectively.
Firms may also choose to delay responding because they understand that many interrelated changes
to a firm’s strategy and operational routines will be needed to respond effectively, and implementing
such comprehensive changes requires time.
H1: Responder performance and response delay are related in an inverted Ushaped manner: Lower
performance will result for fast and late responders, and higher performance for firms with
intermediate response delays.
What the research question is? What is de relatie tussen performance en de delay in het antwoord op
competitie? Eerdere onderzoeken stellen negatieve relatie, maar dit paper stelt een inverted u.
--
In competitive environments, the first mover’s ability to reap the benefits of a move is time
dependent: the later the first moving firm faces competitive responses, the more time it has to
absorb first mover benefits. A fast response is likely to leave the first mover less well off because the
period available for establishing first mover advantages is cut short. This predicted positive linear
relationship for first movers is not affected by response timing errors, that is, responses that are too
slow or too fast and so of less value to the responder. Conscious choices or blind spots that lead to
late responses, for example, simply provide the first mover with more time to claim benefits. If
responders move too fast due to conscious choices, or blind spots, these ill-conceived responses
nevertheless harm the first mover. As fast responses often imitate the original first mover action, they
directly counteract the first mover’s attempts to build market share or achieve other gains.
H2: First mover performance and response delay are related in a linear manner: As response delays
are longer (shorter), performance decreases are smaller (greater).
Context
US retail industry, because characterized by: 1) high level of competitiveness, 2) rich diversity of
competitive interactions, and 3) a broad coverage of competitive moves in the media.
Rivalry in the US retail industry is intense: customer loyalty is weak because buying decisions are
often spontaneous, Walmart’s dominant market position, innovations and continued growth created
competitive pressures on all incumbents, the time frame is 1994 – 2000 in which Internet-based
retailing business models posed additional new threats, and substantial performance variability over
the time of the study was observed.
,Besides, the industry is diverse in strategic and tactical competitive moves and is sufficiently large to
receive daily media and financial attention.
For this study, media attention to the industry was important because identification of competitive
moves relied on this.
Sample
17 of US largest departments and variety stores. Smaller retailers were excluded because of their
limited geographic and competitive scope and limited media attention.
Method
They employed structured content analysis of press articles. One constructs datasets by grouping the
competitive moves occurring within an industry into types and defining a coding scheme to search
electronic databases for reports of these types of actions and responses.
They identified and accessed press articles and newswire reports utilizing Factiva, a division of Dow
Jones & Company that provides research and business information via products (databases) that
access over 10,000 publications.
Competitive dynamics methodology proposes two ways of identifying competitive responses, and
both approaches are used in this study. The original method proposed relies on an explicit reference
to an earlier action in the press article. However, not all articles actually make such references
because at the time of their publication, the original action may have been quite evident to readers
and so required no additional mention. Therefore, an alternative coding approach is to focus on
finding competitive moves that are similar to earlier actions.
This criterion for identifying similarity has two advantages. First, the criterion is stronger than just
comparing moves by type, because in addition to being the same type, two moves also had to focus
on the same product category, geographical market, or customer groups. Second, this criterion
exceeds the same-type imitation logic, because two moves do not need to be of the same type to be
similar.
We defined twelve months as the time frame within which we expected a response to occur if it was
going to occur.
Event study methodology
Suited for this study because it is possible to link reported comp moves to performance effects.
Stock effects were calculated from daily stock return data. A two-day event window was constructed
for every action in the dataset consisting of the day prior to the press release describing the
competitive move and the day of the release itself. The day prior to the press release was included
because information may have already reached the market during this trading day.
Actual stock returns in the event window were compared with hypothetical returns calculated from a
capital asset pricing model that was dynamically estimated for each move using the returns of the
120 days prior to the event window.
This procedure represents the stock effect as a deviation from the underlying market-imputed stock
return in standard deviation units of a standard normal distribution under the null hypothesis that no
stock effect has occurred. This procedure is conservative, as it attributes lower values for the stock
effect of those events that do create additional returns because, in these cases, return variance is
overstated.
What methodology is used? Kwantitatief. Er moet kwalitatieve data geanalyseerd en geclassificeerd
worden, er is ook een deel content analysis. Om de regressie te runnen moet een event study
analysis worden gedaan, dit kijkt naar de impact van een announcement op de stock price voor een
korte tijdsperiode
Dep var: stock effects of responses to capture performance effects for actors and responders.
2 kinds of stock effects were calculated:
, - To determine the wealth created for the responder by his/her response, stock effects were
calculated for the responder at the day of his/her response.
- The wealth effect of the response for the original actor was determined by calculating a stock
effect for the first mover on the day of the response.
Indep var: response delay calculated in working days as the difference between the date of the
action and the date of the response.
In order not to violate regression assumptions, our analysis used the natural logarithm of response
delay.
Control vars
- Move-specific vars: imitation (a response can be an imitation of the original move or it can be
more creative/different – this study does not expect it to matter on their hypotheses) and
tactical actions (Strategic actions such as new product introductions or mergers require
significant investments of specific resources and long time horizons, whereas tactical actions
such as pricing or advertising changes involve fewer and less specific resources and are less
difficult to implement and reverse – this study does not expect it to matter on the
hypotheses).
- Firm-specific vars: firm size, financial performance and age
- Industry-specific vars: industry concentration and industry sales growth
Results
The average response delay was 130 business days for the responder model and 136 business days
for the first mover model. Most responses were imitations with less than one nonimitative response
in five, indicating that many responders intend to copy to catch up, potentially further escalating
competition.
Strategic and tactical actions are balanced, both types of comp moves evoke responses at
approximately equal rates.
Retailers use a broad range of actions to succeed in the marketplace, and need to compete vigorously
to succeed.
Multicollinearity did not pose a problem, as the correlations among the independent variables are
insignificant with three exceptions.
First, response delay and tactical action are negatively correlated, as expected.
Second, industry concentration is positively correlated to imitation, suggesting that changes of
industry market shares that favor the leading firms encourage responders to imitate.
Third, higher industry sales growth is associated with longer response delays, indicating that firms
take more time to respond in industry growth environments.
There is only one significant set of correlations between the dependent and independent variables,
namely a negative correlation between stock effect and response delay in the responder model, and a
positive correlation in the first mover model. These correlations are in line with the linear predictions
of traditional theoretical research with respect to first mover advantages. However, and as our
multivariate analyses below show, they are also misleading.
H1 and H2 are supported.
Limitations
- The questions underlying this research require it to shift from studying the factors that
facilitate fast and aggressive responses to include factors that might cause and prevent errors
when responding quickly. Future research may investigate how conscious choices and blind
spots in competitor analysis lead to premature or late responses
- Response timing may also be influenced by a firm’s resource positions
